Engineering Proceedings

Other

8 pages, 455 KiB

Open AccessProceeding Paper

Explaining When Deep Learning Models Are Better for Time Series Forecasting

by Martín Solís and Luis-Alexander Calvo-Valverde

Eng. Proc. 2024, 68(1), 1; https://doi.org/10.3390/engproc2024068001 - 27 Jun 2024

Viewed by 481

There is a gap of knowledge about the conditions that explain why a method has a better forecasting performance than another. Specifically, this research aims to find the factors that can influence deep learning models to work better with time series. We generated [...] Read more.

There is a gap of knowledge about the conditions that explain why a method has a better forecasting performance than another. Specifically, this research aims to find the factors that can influence deep learning models to work better with time series. We generated linear regression models to analyze if 11 time series characteristics influence the performance of deep learning models versus statistical models and other machine learning models. For the analyses, 2000 time series of M4 competition were selected. The results show findings that can help explain better why a pretrained deep learning model is better than another kind of model. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 1006 KiB

Open AccessProceeding Paper

The Cassandra Method: Dystopian Visions as a Basis for Responsible Design

by Sarah Diefenbach and Daniel Ullrich

Eng. Proc. 2024, 68(1), 2; https://doi.org/10.3390/engproc2024068002 - 27 Jun 2024

Viewed by 291

Abstract

Innovative technologies often have unforeseen negative consequences on an individual, societal, or environmental level. To minimize these, the Cassandra method aims to foresee such negative effects by systematically investigating dystopian visions. Starting with the activation of a (self-)critical mindset, the next steps are [...] Read more.

Innovative technologies often have unforeseen negative consequences on an individual, societal, or environmental level. To minimize these, the Cassandra method aims to foresee such negative effects by systematically investigating dystopian visions. Starting with the activation of a (self-)critical mindset, the next steps are collecting a maximum number of negative effects and assessing their relevance. Finally, the envisioned impairments are used to improve the product concepts in a responsible way. This paper broadly outlines the method, its applications during product development and research, and reports on experiences from an expert workshop. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 571 KiB

Open AccessProceeding Paper

Forecasting Methods for Road Accidents in the Case of Bucharest City

by Cristina Oprea, Eugen Rosca, Ionuț Preda, Anamaria Ilie, Mircea Rosca and Florin Rusca

Eng. Proc. 2024, 68(1), 3; https://doi.org/10.3390/engproc2024068003 - 27 Jun 2024

Viewed by 386

Abstract

This paper aims to emphasize the necessity for policy reform, improvements in vehicle design and enhanced public awareness through the projection of future trends in road accidents, injuries and fatalities. The statistical methods that are used in this study are the empirical laws [...] Read more.

This paper aims to emphasize the necessity for policy reform, improvements in vehicle design and enhanced public awareness through the projection of future trends in road accidents, injuries and fatalities. The statistical methods that are used in this study are the empirical laws of Smeed and Andreassen. The main gap that the researchers identify is the lack of a standardized methodology with the help of which the appropriate forecasting method can be chosen in the area of traffic accidents. In the present study, the authors propose such a methodology that can be generalized, being suitable for use for any urban agglomeration at the micro and macro level. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 880 KiB

Open AccessProceeding Paper

Deep Learning for Crime Forecasting of Multiple Regions, Considering Spatial–Temporal Correlations between Regions

by Martín Solís and Luis-Alexander Calvo-Valverde

Eng. Proc. 2024, 68(1), 4; https://doi.org/10.3390/engproc2024068004 - 28 Jun 2024

Viewed by 395

Abstract

Crime forecasting has gained popularity in recent years; however, the majority of studies have been conducted in the United States, which may result in a bias towards areas with a substantial population. In this study, we generated different models capable of forecasting the [...] Read more.

Crime forecasting has gained popularity in recent years; however, the majority of studies have been conducted in the United States, which may result in a bias towards areas with a substantial population. In this study, we generated different models capable of forecasting the number of crimes in 83 regions of Costa Rica. These models include the spatial–temporal correlation between regions. The findings indicate that the architecture based on an LSTM encoder–decoder achieved superior performance. The best model achieved the best performance in regions where crimes occurred more frequently; however, in more secure regions, the performance decayed. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 282 KiB

Open AccessProceeding Paper

Optimizing Social Security Contributions for Spanish Self-Employed Workers: Combining Data Preprocessing and Ensemble Models for Accurate Revenue Estimation

by Luis Palomero, Vicente García and José Salvador Sánchez

Eng. Proc. 2024, 68(1), 5; https://doi.org/10.3390/engproc2024068005 - 28 Jun 2024

Viewed by 354

Abstract

The Real Decreto-ley 13/2022 has amended the framework governing the calculation of Social Security contributions for Spanish self-employed workers. This framework obligates taxpayers to the annual revenue projection, under the possibility of lending money for free or paying unexpected taxes at the end [...] Read more.

The Real Decreto-ley 13/2022 has amended the framework governing the calculation of Social Security contributions for Spanish self-employed workers. This framework obligates taxpayers to the annual revenue projection, under the possibility of lending money for free or paying unexpected taxes at the end of the year in the case of deviations. To address this issue, the Declarando firm has developed an algorithm to recommend the optimal contributions that combines a Simple Moving Average forecasting method with an offset-adjustment technique. This paper examines how this strategy can be improved by cleaning the input data and combining different forecasts using an Ensemble-based approach. After testing experimentally various alternatives, a promising strategy involves employing a median-based Ensemble on preprocessed data. Although this Ensemble-based approach significantly reduces forecasting errors, the improvements are diluted when the predictions are combined with the offset-adjustment process. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 668 KiB

Open AccessProceeding Paper

Studying LF and HF Time Series to Characterize Cardiac Physiological Responses to Mental Fatigue

by Alexis Boffet, Veronique Deschodt Arsac and Eric Grivel

Eng. Proc. 2024, 68(1), 6; https://doi.org/10.3390/engproc2024068006 - 28 Jun 2024

Viewed by 324

Abstract

Heart rate variability (HRV) was largely used to evaluate psychophysiological status of Human at rest as well as during cognitive tasks, for both healthy subjects and patients. Among the approaches used for assessing cardiac autonomic control from HRV analysis, biomarkers such as the [...] Read more.

Heart rate variability (HRV) was largely used to evaluate psychophysiological status of Human at rest as well as during cognitive tasks, for both healthy subjects and patients. Among the approaches used for assessing cardiac autonomic control from HRV analysis, biomarkers such as the power in low and high frequencies (LF-HF) are often extracted from short-term recordings lasting 2 to 5 min. Although they correctly reflect the average psychophysiological state of a subject in situation, they fail to analyse cardiac autonomic control over time. For this reason, we suggest investigating the LF-HF biomarkers over time to identify mental fatigue and determine different physiological profiles. The following step consists in defining the set of parameters that characterise the LF-HF time series and that can be interpreted easily by the physiologists. In this work, polynomial models are considered to describe the trends of the LF-HF time series. The latter are then decomposed into decreasing (d) and increasing (i) parts. Finally, the proportion of the i parts of the polynomial trends of the LF and HF powers over time are combined with classically-used metrics to define individual profiles in response to mental fatigue. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Experimental protocol. INITIAL: Mental Fatigue perceived (VAS); FINAL: Mental Fatigue Perceived (VAS) and Workload perceived (NASA_TLX). Full article ">Figure 2
Illustration of the approach with patient 29. The HF process (in blue) modelled by a 5th-degree polynomial (in red) and resulting signed parameters: <math display="inline"><semantics> <mrow> <mi>i</mi> <mi>H</mi> <mi>F</mi> </mrow> </semantics></math> = 0.311. The sign − is added before the value of the percentage to express the fact that it is a decreasing trend. Full article ">Figure 3
Evolution of perceived mental fatigue before and after CVT. Score 1: “absence of mental fatigue”; Score 10: “extreme level of mental fatigue”; ***: p-value < 0.001. Full article ">Figure 4
K-means clustering performed on a dataset of 42 volunteers from the average of RR, RMSSD, LF power, HF power, LF/HF ratio, and RCMSE calculated over the first 10 min and last 10 min of CVT, combined with the time course markers <math display="inline"><semantics> <mrow> <mi>i</mi> <mi>L</mi> <mi>F</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <mi>i</mi> <mi>H</mi> <mi>F</mi> </mrow> </semantics></math>. For the sake of simplicity, we represent the clusters with their respective characteristics with respect to each dimension (Dim1 and Dim2). Full article ">

9 pages, 11746 KiB

Open AccessProceeding Paper

Annual Runoff Forecasting through Bayesian Causality

by Santiago Zazo, Jose-Luis Molina, Carmen Patino-Alonso, Fernando Espejo and Juan Carlos García-Prieto

Eng. Proc. 2024, 68(1), 7; https://doi.org/10.3390/engproc2024068007 - 28 Jun 2024

Viewed by 292

Abstract

This contribution is focused on forecasting ability of Bayesian causality (BC) on annual runoff series. For that, the time series was synthetized through a Bayesian net, in which the probability propagation over the time was performed. The BC analytical ability identified the hidden [...] Read more.

This contribution is focused on forecasting ability of Bayesian causality (BC) on annual runoff series. For that, the time series was synthetized through a Bayesian net, in which the probability propagation over the time was performed. The BC analytical ability identified the hidden logic structure of the hydrological records that describes its overall behavior. This allowed us to quantify the runoff, through a novel dependence matrix, through two fractions, one conditional on time (temporally conditioned runoff) and one not (temporally nonconditioned runoff). This conditionality allowed the development of two predictive models for each fraction, analyzing their reliability under a double probabilistic and metrological approach. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 2414 KiB

Open AccessProceeding Paper

Towards Resolving the Ambiguity in Low-Field, All-Optical Magnetic Field Sensing with High NV-Density Diamonds

by Ludwig Horsthemke, Jens Pogorzelski, Dennis Stiegekötter, Frederik Hoffmann, Ann-Sophie Bülter, Sarah Trinschek, Markus Gregor and Peter Glösekötter

Eng. Proc. 2024, 68(1), 8; https://doi.org/10.3390/engproc2024068008 - 1 Jul 2024

Viewed by 411

Abstract

In all-optical magnetic field sensing using nitrogen-vacancy-center-rich diamonds, an ambiguity in the range of 0–8 mT can be observed. We propose a way to resolve this ambiguity using the magnetic-field-dependent fluorescence lifetime. We therefore recorded the frequency response of the fluorescence upon modulation [...] Read more.

In all-optical magnetic field sensing using nitrogen-vacancy-center-rich diamonds, an ambiguity in the range of 0–8 mT can be observed. We propose a way to resolve this ambiguity using the magnetic-field-dependent fluorescence lifetime. We therefore recorded the frequency response of the fluorescence upon modulation of the excitation intensity in a frequency range of 1–

100 MHz

. The magnetic-field-dependent decay dynamics led to different response characteristics for magnetic fields below and above

3 mT

, allowing us to resolve the ambiguity. We used a physics-based model function to extract fit parameters, which we used for regression, and compared it to an alternative approach purely based on an artificial neural network. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 942 KiB

Open AccessProceeding Paper

Modeling a Set of Variables with Different Attributes on a Quantitative Dependent Variable: An Application of Dichotomous Variables

by Gerardo Covarrubias and Xuedong Liu

Eng. Proc. 2024, 68(1), 9; https://doi.org/10.3390/engproc2024068009 - 1 Jul 2024

Viewed by 271

Abstract

This study outlines the methodology employed to model the relationship among a set of dichotomous variables, which represent attributes, on a nominal scale. The objective is to elucidate their influence on a quantitative dependent variable measured on a ratio scale. This approach allows [...] Read more.

This study outlines the methodology employed to model the relationship among a set of dichotomous variables, which represent attributes, on a nominal scale. The objective is to elucidate their influence on a quantitative dependent variable measured on a ratio scale. This approach allows for the quantification of the impact of these attributes and their significance in shaping the behavior of the entity possessing them. The resolution method employed for the estimation is ordinary least squares. However, it is crucial to note that interpreting the estimators in the resulting model requires a nuanced perspective, distinguishing it from the conventional interpretation of slope or rate of change in a classic model. To clarify, these estimators align with the average behavior of the dependent variable concerning binary characteristics, and the outcomes are consistent with the analysis of variance. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1495 KiB

Open AccessProceeding Paper

A Python Module for Implementing Cointegration Tests with Multiple Endogenous Structural Breaks

by Abdulnasser Hatemi-J and Alan Mustafa

Eng. Proc. 2024, 68(1), 10; https://doi.org/10.3390/engproc2024068010 - 2 Jul 2024

Viewed by 375

Abstract

Testing for long-run relationships between time series variables with short-run adjustments is an integral part of many empirical studies nowadays. Allowing for structural breaks in the estimations is a pertinent issue within this context. The purpose of this paper is to provide a [...] Read more.

Testing for long-run relationships between time series variables with short-run adjustments is an integral part of many empirical studies nowadays. Allowing for structural breaks in the estimations is a pertinent issue within this context. The purpose of this paper is to provide a consumer-friendly module that is created in Python for implementing three residuals-based cointegration tests with two unknown regime shifts. The timing of each shift is revealed endogenously. The software is easy to use via a Graphical User Interface (GUI). In addition to implementing cointegration tests, the software also estimates the underlying parameters along with the standard errors and the significance tests for the parameters. An application is also provided using real data to demonstrate how the software can be used. To our best knowledge, this is the first software component created in Python that implements cointegration tests with structural breaks. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1592 KiB

Open AccessProceeding Paper

Big Data Techniques Applied to Forecast Photovoltaic Energy Demand in Spain

by J. Tapia-García, L. G. B. Ruiz, D. Criado-Ramón and M. C. Pegalajar

Eng. Proc. 2024, 68(1), 11; https://doi.org/10.3390/engproc2024068011 - 3 Jul 2024

Viewed by 365

Abstract

Renewable energies play an important role in our society’s development, addressing the challenges presented by climate change. Specifically, in countries like Spain, technologies such as solar energy assume a crucial significance, enabling the generation of clean energy. This study addresses the critical need [...] Read more.

Renewable energies play an important role in our society’s development, addressing the challenges presented by climate change. Specifically, in countries like Spain, technologies such as solar energy assume a crucial significance, enabling the generation of clean energy. This study addresses the critical need to accurately predict photovoltaic (PV) energy demand in Spain. By using the data collected from the Spanish Electricity System, four models (Linear Regression, Random Forest, Recurrent Neural Network, and LightGBM) were implemented, with adaptations for Big Data. The LR model proved unsuitable, while the LGBM emerged as the most accurate and timely performer. The incorporation of Big Data adaptations amplifies the significance of our findings, highlighting the effectiveness of the LGBM in forecasting PV energy demand with both accuracy and efficiency. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 6368 KiB

Open AccessProceeding Paper

Detecting Trend Turning Points in PS-InSAR Time Series: Slow-Moving Landslides in Province of Frosinone, Italy

by Ebrahim Ghaderpour, Benedetta Antonielli, Francesca Bozzano, Gabriele Scarascia Mugnozza and Paolo Mazzanti

Eng. Proc. 2024, 68(1), 12; https://doi.org/10.3390/engproc2024068012 - 3 Jul 2024

Cited by 1 | Viewed by 361

Abstract

Detecting slow-moving landslides is a crucial task for mitigating potential risk to human lives and infrastructures. In this research, Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) time series, provided by the European Ground Motion Service (EGMS), for the province of Frosinone in Italy [...] Read more.

Detecting slow-moving landslides is a crucial task for mitigating potential risk to human lives and infrastructures. In this research, Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) time series, provided by the European Ground Motion Service (EGMS), for the province of Frosinone in Italy are employed, and Sequential Turning Point Detection (STPD) is applied to them to estimate the dates when the displacement rates change. The estimated dates are classified based on the land cover/use of the province. Moreover, local precipitation time series are employed to investigate how precipitation rate changes might have triggered the landslides. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 2175 KiB

Open AccessProceeding Paper

Measuring the Efficiency of Introducing Businesses’ Digitalization Elements over Time in Relation to Their Performance

by Jarmila Horváthová and Martina Mokrišová

Eng. Proc. 2024, 68(1), 13; https://doi.org/10.3390/engproc2024068013 - 3 Jul 2024

Viewed by 341

Abstract

The introduction of digitalization elements into the life of companies is significant in terms of achieving better economic results. The aim of the research was to determine the technical efficiency, as well as the change in efficiency and the technological change in the [...] Read more.

The introduction of digitalization elements into the life of companies is significant in terms of achieving better economic results. The aim of the research was to determine the technical efficiency, as well as the change in efficiency and the technological change in the digital transformation of companies in EU countries in relation to their performance. The Malmquist index was used to measure these parameters over time. The results of the research indicate the significance of the dynamic measurement of the efficiency of digital transformation. Interesting results also point to the importance of evaluating the efficiency of the use of already established elements, as well as evaluating the introduction of new technological changes. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 915 KiB

Open AccessProceeding Paper

Evaluation of the University of Lagos Waste Generation Trend

by Charles A. Mbama, Austin Otegbulu, Iain Beverland and Tara K. Beattie

Eng. Proc. 2024, 68(1), 14; https://doi.org/10.3390/engproc2024068014 - 4 Jul 2024

Viewed by 440

Abstract

This study examines waste generation patterns at the University of Lagos (UoL), Nigeria, to inform decision-making towards improving the efficiency of the university’s management strategies in line with Sustainable Development Goal 12, target 12.5 to reduce waste generation through prevention, reduction, recycling, and [...] Read more.

This study examines waste generation patterns at the University of Lagos (UoL), Nigeria, to inform decision-making towards improving the efficiency of the university’s management strategies in line with Sustainable Development Goal 12, target 12.5 to reduce waste generation through prevention, reduction, recycling, and reuse by 2030. The moving average of the waste generation was studied using time series data. During October 2014 to October 2016 the UoL generated an average of 877.5 tons of waste every month, with the lowest observed value being 496.6 tons and the highest recorded value being 1250.5 tons. The trend result indicates a gradual decrease in the generation of waste over time. There is also a noticeable negative cyclical pattern with seasonal variations, where the highest generation point is observed in March and the lowest point is observed in June, particularly in the latter half of the second quarter, as time progresses. Although there is a reduction in the amount of waste generated over time, it is crucial to persist in evaluating diverse waste management strategies that could further reduce the amount of waste generated in the case study area. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 834 KiB

Open AccessProceeding Paper

Modeling the Asymmetric and Time-Dependent Volatility of Bitcoin: An Alternative Approach

by Abdulnasser Hatemi-J

Eng. Proc. 2024, 68(1), 15; https://doi.org/10.3390/engproc2024068015 - 4 Jul 2024

Viewed by 476

Abstract

Volatility as a measure of financial risk is a crucial input for hedging, portfolio diversification, option pricing and the calculation of the value at risk. In this paper, we estimate the asymmetric and time-varying volatility for Bitcoin as the dominant cryptocurrency in the [...] Read more.

Volatility as a measure of financial risk is a crucial input for hedging, portfolio diversification, option pricing and the calculation of the value at risk. In this paper, we estimate the asymmetric and time-varying volatility for Bitcoin as the dominant cryptocurrency in the world market. A novel approach that explicitly separates the falling markets from the rising ones is utilized for this purpose. The empirical results have important implications for investors and financial institutions. Our approach provides a position-dependent measure of risk for Bitcoin. This is essential since the source of risk for an investor with a long position is the falling prices, while the source of risk for an investor with a short position is the rising prices. Thus, providing a separate risk measure in each case is expected to increase the efficiency of the underlying risk management in both cases compared to the existing methods in the literature. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 1376 KiB

Open AccessProceeding Paper

Modelling the Daily Concentration of Airborne Particles Using 1D Convolutional Neural Networks

by Ivan Gudelj, Mario Lovrić and Emmanuel Karlo Nyarko

Eng. Proc. 2024, 68(1), 16; https://doi.org/10.3390/engproc2024068016 - 4 Jul 2024

Viewed by 384

Abstract

This paper focuses on improving the prediction of the daily concentration of the pollutants, PM₁₀ and nitrogen oxides (NO, NO₂) in the air at urban monitoring sites using 1D convolutional neural networks (CNN). The results show that the 1D CNN [...] Read more.

This paper focuses on improving the prediction of the daily concentration of the pollutants, PM₁₀ and nitrogen oxides (NO, NO₂) in the air at urban monitoring sites using 1D convolutional neural networks (CNN). The results show that the 1D CNN model outperforms the other machine learning models (LSTM and Random Forest) in terms of the coefficients of determination and absolute errors. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

16 pages, 2534 KiB

Open AccessProceeding Paper

Reservoir Neural Network Computing for Time Series Forecasting in Aerospace: Potential Applications to Predictive Maintenance

by Juan Manuel Rodríguez Riesgo and Juan Luis Cabrera Fernández

Eng. Proc. 2024, 68(1), 17; https://doi.org/10.3390/engproc2024068017 - 4 Jul 2024

Viewed by 430

Abstract

Coupling a reservoir neural network and a Grey Wolf optimization algorithm the system hyperparameters space is explored to find the configuration best suited to forecast the input sensor from the NASA CMAPSS dataset. In such a framework, the application to the problem of [...] Read more.

Coupling a reservoir neural network and a Grey Wolf optimization algorithm the system hyperparameters space is explored to find the configuration best suited to forecast the input sensor from the NASA CMAPSS dataset. In such a framework, the application to the problem of predictive maintenance is considered. The necessary requirements for the system to generate satisfactory predictions are established, with specific suggestions as to how a forecast can be improved through reservoir computing. The obtained results are used to determine certain common rules that improve the quality of the predictions and focus the optimization towards hyperparameter solutions that may allow for a faster approach to predictive maintenance. This research is a starting point to develop methods that could inform accurately on the remaining useful life of a component in aerospace systems. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
RC-GWO system flowchart. The sensor data enters as an input for pre-processing and hyperparameters optimization to find a compatible reservoir to generate the prediction upon. Full article ">Figure 2
CMAPSS dataset Test 1, Unit 49 (from top to bottom): OS 1: Altitude, OS 2: Mach Number, Sensor 2: Total temperature at LPC outlet, Sensor 7: Total pressure at HPC outlet. Time units are in operating cycles. Full article ">Figure 3
Internal states of a reservoir. Left: Internal states for each the neuron. Right: histogram for the internal states. Results obtained at an arbitrary step during training. Top: A saturated reservoir. Bottom: an unsaturated reservoir. Full article ">Figure 4
CMAPSS dataset Test 1, Unit 49 time series prediction for Set 3 (from top to bottom): OS 1: Altitude, OS 2: Mach Number, Sensor 2: Total temperature at LPC outlet and, Sensor 7: total pressure at HPC outlet. Time units are in operating cycles. Green: target signal; Blue: predicted signal. Results for the additional sets are available as <a href="#app1-engproc-68-00017" class="html-app">Supplementary Materials</a>. Full article ">Figure 5
Zoom detail of the forecasting along the 10 initial cycles. Details are the same as in <a href="#engproc-68-00017-f004" class="html-fig">Figure 4</a>. Full article ">

8 pages, 278 KiB

Open AccessProceeding Paper

Comparison of Inferential Methods for a Novel CMP Model

by Yuvraj Sunecher and Naushad Mamode Khan

Eng. Proc. 2024, 68(1), 18; https://doi.org/10.3390/engproc2024068018 - 4 Jul 2024

Viewed by 205

Abstract

In many real-life instances, time series of counts are often exposed to the dispersion phenomenon while at the same time being influenced by some explanatory variables. This paper takes into account these two issues by assuming that the series of counts follow an [...] Read more.

In many real-life instances, time series of counts are often exposed to the dispersion phenomenon while at the same time being influenced by some explanatory variables. This paper takes into account these two issues by assuming that the series of counts follow an observation-driven first-order integer-valued moving-average structure (INMA(1)) where the innovation terms are COM-Poisson (CMP) distributed under a link function characterized by time-independent covariates. The second part of the paper constitutes estimating the regression effects, dispersion and the serial parameters using popular estimation methods, mainly the Conditional Least Squares (CLS), Generalized Method of Moments and Generalized Quasi-Likelihood (GQL) approaches. The performance of these estimation methods is compared via simulation experiments under different levels of dispersion. Additionally, the suggested model is used to examine time series accident data from actual incidents in several Mauritius locations. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

7 pages, 730 KiB

Open AccessProceeding Paper

Foreign Exchange Forecasting Models: LSTM and BiLSTM Comparison

by Fernando García, Francisco Guijarro, Javier Oliver and Rima Tamošiūnienė

Eng. Proc. 2024, 68(1), 19; https://doi.org/10.3390/engproc2024068019 - 4 Jul 2024

Viewed by 359

Abstract

Knowledge of foreign exchange rates and their evolution is fundamental to firms and investors, both for hedging exchange rate risk and for investment and trading. The ARIMA model has been one of the most widely used methodologies for time series forecasting. Nowadays, neural [...] Read more.

Knowledge of foreign exchange rates and their evolution is fundamental to firms and investors, both for hedging exchange rate risk and for investment and trading. The ARIMA model has been one of the most widely used methodologies for time series forecasting. Nowadays, neural networks have surpassed this methodology in many aspects. For short-term stock price prediction, neural networks in general and recurrent neural networks such as the long short-term memory (LSTM) network in particular perform better than classical econometric models. This study presents a comparative analysis between the LSTM model and BiLSTM models. There is evidence for an improvement in the bidirectional model for predicting foreign exchange rates. In this case, we analyse whether this efficiency is consistent in predicting different currencies as well as the bitcoin futures contract. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 336 KiB

Open AccessProceeding Paper

Extraction and Forecasting of Trends in Cases of Signal Rank Overestimation

by Nina Golyandina and Pavel Dudnik

Eng. Proc. 2024, 68(1), 20; https://doi.org/10.3390/engproc2024068020 - 5 Jul 2024

Viewed by 257

Abstract

Singular spectrum analysis allows automation for the extraction of trends of arbitrary shapes. Here, we study how the estimation of signal ranks influences the accuracy of trend extraction and forecasting. It is numerically shown that the trend estimates and their forecasting are slightly [...] Read more.

Singular spectrum analysis allows automation for the extraction of trends of arbitrary shapes. Here, we study how the estimation of signal ranks influences the accuracy of trend extraction and forecasting. It is numerically shown that the trend estimates and their forecasting are slightly changed if the signal rank is overestimated. If the trend is not of finite rank, the trend estimates are still stable, while forecasting may be unstable. The method of automation of the trend extraction includes an important step for improving the separability of time series components to avoid their mixture. However, the better the separability improvement, the larger the forecasting variability, since the noise components become separated and can be similar to trend ones. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 2110 KiB

Open AccessProceeding Paper

Forecasting Stock Market Dynamics using Market Cap Time Series of Firms and Fluctuating Selection

by Hugo Fort

Eng. Proc. 2024, 68(1), 21; https://doi.org/10.3390/engproc2024068021 - 5 Jul 2024

Viewed by 806

Abstract

Evolutionary economics has been instrumental in explaining the nature of innovation processes and providing valuable heuristics for applied research. However, quantitative tests in this field remain scarce. A significant challenge is accurately estimating the fitness of companies. We propose the estimation of the [...] Read more.

Evolutionary economics has been instrumental in explaining the nature of innovation processes and providing valuable heuristics for applied research. However, quantitative tests in this field remain scarce. A significant challenge is accurately estimating the fitness of companies. We propose the estimation of the financial fitness of a company by its market capitalization (MC) time series using Malthusian fitness and the selection equation of evolutionary biology. This definition of fitness implies that all companies, regardless of their industry, compete for investors’ money through their stocks. The resulting fluctuating selection from market capitalization (FSMC) formula allows forecasting companies’ shares of total MC through this selection equation. We validate the method using the daily MC of public-owned Fortune 100 companies over the period 2000–2021. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Estimation of fitness: instantaneous vs. smoothed fitness. Data corresponding to Apple (AAPL) for the second and third quarters of 2021. The rapidly varying black full line is the instantaneous fitness produced by Equation (6) for each day of the validation period. The thick gray line is the smoothed fitness, obtained through Equation (7) with a running time window of length TT = 63 days (the number of market days of a quarter); it shows much smaller variations and slightly departs from the constant fitness value (red-dotted line) used by the FSMC forecasting. Full article ">Figure 2
The absolute error of FSMC predictions for each firm over Tv = 252 days (a market year). Each of the 252 × 78 cells corresponds to the absolute error for the forecasted day and company number i averaged over the 5536 − 2TV = 5536 − 2 × 252 = 5032 validation instances (Equation (10)). The color code is as follows: blue indicates errors smaller than the value of the mean fractions. mean{xi(t)} = 0.025. Full article ">Figure 3
The absolute percentage errors yielded by the FSMC method for each firm over Tv = 21 days (a market month). Each of the 21 × 78 cells corresponds to the percentage error for the forecasted day and company number i averaged over the 5536 − 2TV = 5536 − 2 × 21 = 5494 validation instances (Equation (11)). The color code is as follows: blue indicates small average relative errors (<5%), while red corresponds to large relative errors (>20%). Full article ">Figure 4
MAPE of FSMC forecast (Equation (13)) over TV = 21 days (a month) for each firm. Full article ">Figure 5
The evolution of the market caps of AIG and Fannie Mae from 2000 to 2021. The inset is zoomed in on the corresponding fractions of both companies (filled) and the FSMC predictions (dashed and dotted). $ corresponds to USD. Full article ">

7 pages, 812 KiB

Open AccessProceeding Paper

Using Dichotomous Variables to Model Structural Changes in Time Series: An Application to International Trade

by Gerardo Covarrubias and Xuedong Liu

Eng. Proc. 2024, 68(1), 22; https://doi.org/10.3390/engproc2024068022 - 5 Jul 2024

Viewed by 239

Abstract

This research aimed to elucidate the methodology employed in econometric estimations by utilizing dichotomous variables. These variables served a dual purpose: firstly, they denoted an attribute designed to discern structural changes within a linear relationship, and secondly, they quantified the impact and statistical [...] Read more.

This research aimed to elucidate the methodology employed in econometric estimations by utilizing dichotomous variables. These variables served a dual purpose: firstly, they denoted an attribute designed to discern structural changes within a linear relationship, and secondly, they quantified the impact and statistical significance of one or more quantitative independent variables on a dependent variable in the presence of such structural changes. The core of this document covers the presentation of the methodology; additionally, an applied analysis of Mexico’s foreign trade is included. This analysis delved into estimating the impact and statistical significance of exports on the economic growth across different periods, reflecting significant structural changes in the development of the Mexican economy. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 4292 KiB

Open AccessProceeding Paper

A Global Deep Learning Perspective on Australia-Wide Monthly Precipitation Prediction

by Luyi Shen, Guoqi Qian and Antoinette Tordesillas

Eng. Proc. 2024, 68(1), 23; https://doi.org/10.3390/engproc2024068023 - 8 Jul 2024

Viewed by 281

Abstract

Gaining a deep understanding of precipitation patterns is beneficial for enhancing Australia’s adaptability to climate change. Driven by this motivation, we present a specific spatiotemporal deep learning model that well integrates matrix factorization and temporal convolutional networks, along with essential year-month covariates and [...] Read more.

Gaining a deep understanding of precipitation patterns is beneficial for enhancing Australia’s adaptability to climate change. Driven by this motivation, we present a specific spatiotemporal deep learning model that well integrates matrix factorization and temporal convolutional networks, along with essential year-month covariates and key climatic drivers, to analyze and forecast monthly precipitation in Australia. We name this the spatiotemporal TCN-MF method. Our approach employs the precipitation profiler-observation fusion and estimation (PPrOFusE) method for data input, synthesizing monthly precipitation readings from the gauge measurement of the Bureau of Meteorology (BoM), the JAXA Global Satellite Mapping of Precipitation (GSMaP), and the NOAA Climate Prediction Center Morphing (CMORPH) technique. The input dataset spans from April 2000 to March 2021 and covers 1391 Australian grid locations. To evaluate the model’s effectiveness, particularly in regions prone to severe flooding, we employ the empirical dynamic quantiles (EDQ) technique. This method ranks cumulative rainfall levels, enabling focused analysis on areas most affected by extreme weather events. Our assessment from April 2021 to March 2022 highlights the model’s proficiency in identifying significant rainfall, especially in flood-impacted locations. Through the analysis across various climatic zones, the spatiotemporal TCN-MF model contributes to the field of continent-wide precipitation forecasting, providing valuable insights that may enhance climate change adaptability strategies in Australia. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
(a) Precipitation data for Australia in March 2022. From top to bottom: Fused, BoM, JAXA, and NOAA data. (b) Time series representation of climate drivers from April 2000 to March 2022. From top to bottom: Dipole Mode Index (DMI), Southern Annular Mode (SAM), and Southern Oscillation Index (SOI). Full article ">Figure 2
Architecture of the spatiotemporal TCN-MF model. Left: The input and the techniques required for the global model, namely TCN and MF. The base TCN diagram is sourced from [<a href="#B26-engproc-68-00023" class="html-bibr">26</a>]. Right: The local TCN model, integrating transformed precipitation, seasonal and climate covariates, and global model outputs, is shown. By integrating both local and global models, a unique model is executed for each location. Consequently, the final output presents a holistic precipitation forecast across Australia. Full article ">Figure 3
Overview of precipitation data for 1391 locations from April 2021 to March 2022. (Top) Fused precipitation data using the PPrOFusE method, representing observed rainfall. (Bottom) Monthly precipitation forecasts using the spatiotemporal TCN-MF Model. Full article ">Figure 3 Cont.
Overview of precipitation data for 1391 locations from April 2021 to March 2022. (Top) Fused precipitation data using the PPrOFusE method, representing observed rainfall. (Bottom) Monthly precipitation forecasts using the spatiotemporal TCN-MF Model. Full article ">Figure 4
Central heatmap illustrating the ranking of 1391 cumulative monthly precipitation time series across Australia using the EDQ method. Surrounding line graphs display comparative analysis at eight selected locations, with the quantiles corresponding to <math display="inline"><semantics> <mrow> <mo>{</mo> <mn>0.001</mn> <mo>,</mo> <mn>0.503</mn> <mo>,</mo> <mn>0.953</mn> <mo>,</mo> <mn>0.965</mn> <mo>,</mo> <mn>0.971</mn> <mo>,</mo> <mn>0.983</mn> <mo>,</mo> <mn>0.991</mn> <mo>,</mo> <mn>0.996</mn> <mo>}</mo> </mrow> </semantics></math>. Observed rainfall is shown in black, while blue dashed lines represent forecasts from a seasonal VAR model, and red dashed lines indicate predictions from the spatiotemporal TCN-MF model. The x-axis spans the prediction period from April 2021 to March 2022. Full article ">

10 pages, 7610 KiB

Open AccessProceeding Paper

Prediction of the Characteristics of Concrete Containing Crushed Brick Aggregate

by Marijana Hadzima-Nyarko, Miljan Kovačević, Ivanka Netinger Grubeša and Silva Lozančić

Eng. Proc. 2024, 68(1), 24; https://doi.org/10.3390/engproc2024068024 - 8 Jul 2024

Viewed by 299

Abstract

The construction industry faces the challenge of conserving natural resources while maintaining environmental sustainability. This study investigates the feasibility of using recycled materials, particularly crushed clay bricks, as replacements for conventional aggregates in concrete. The research aims to optimize the performance of both [...] Read more.

The construction industry faces the challenge of conserving natural resources while maintaining environmental sustainability. This study investigates the feasibility of using recycled materials, particularly crushed clay bricks, as replacements for conventional aggregates in concrete. The research aims to optimize the performance of both single regression tree models and ensembles of regression trees in predicting concrete properties. The study focuses on optimizing key parameters like the minimum leaf size in the models. By testing various minimum leaf sizes and ensemble methods such as Random Forest and TreeBagger, the study evaluates metrics including Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The analysis indicates that the most influential factors on concrete characteristics are the concrete’s age, the amount of superplasticizer used, and the size of crushed brick particles exceeding 4 mm. Additionally, the water-to-cement ratio significantly impacts the predictions. The regression tree models showed optimal performance with a minimum leaf size, achieving an RMSE of 4.00, an MAE of 2.95, an MAPE of 0.10, and an R² of 0.96. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 1234 KiB

Open AccessProceeding Paper

Enabling Diffusion Model for Conditioned Time Series Generation

by Frédéric Montet, Benjamin Pasquier, Beat Wolf and Jean Hennebert

Eng. Proc. 2024, 68(1), 25; https://doi.org/10.3390/engproc2024068025 - 8 Jul 2024

Viewed by 483

Abstract

Synthetic time series generation is an emerging field of study in the broad spectrum of data science, addressing critical needs in diverse fields such as finance, meteorology, and healthcare. In recent years, diffusion methods have shown impressive results for image synthesis thanks to [...] Read more.

Synthetic time series generation is an emerging field of study in the broad spectrum of data science, addressing critical needs in diverse fields such as finance, meteorology, and healthcare. In recent years, diffusion methods have shown impressive results for image synthesis thanks to models such as Stable Diffusion and DALL·E, defining the new state-of-the-art methods. In time series generation, their potential exists but remains largely unexplored. In this work, we demonstrate the applicability and suitability of diffusion methods for time series generation on several datasets with a rigorous evaluation procedure. Our proposal, inspired from an existing diffusion model, obtained a better performance than a reference model based on generative adversarial networks (GANs). We also propose a modification of the model to allow for guiding the generation with respect to conditioning variables. This conditioned generation is successfully demonstrated on meteorological data. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1557 KiB

Open AccessProceeding Paper

Estimation and Prediction of Cereal Production Using Normalized Difference Vegetation Index Time Series (Sentinel-2) Data in Central Spain

by César Sáenz, Alfonso Bermejo-Saiz, Víctor Cicuéndez, Tomás Pugni, Diego Madruga, Alicia Palacios-Orueta and Javier Litago

Eng. Proc. 2024, 68(1), 26; https://doi.org/10.3390/engproc2024068026 - 8 Jul 2024

Viewed by 320

Abstract

Estimating production in cereal fields allows farmers to obtain information on improving management in their following campaigns and avoiding losses. The main objective of this work was to estimate grain production in cereals (wheat and barley) in the 2019 and 2020 campaigns in [...] Read more.

Estimating production in cereal fields allows farmers to obtain information on improving management in their following campaigns and avoiding losses. The main objective of this work was to estimate grain production in cereals (wheat and barley) in the 2019 and 2020 campaigns in three provinces of Central Spain. The model was based on the prediction of the maximum values of the Sentinel-2 Normalized Difference Vegetation Index (NDVI) time series with ARIMA and multiple linear regression models. The highest correlation was found between grain yield and the variables’ five-month cumulative rainfall and maximum greenness (NDVI_max). Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 1594 KiB

Open AccessProceeding Paper

Exploring Optimal Strategies for Small Hydro Power Forecasting: Training Periods and Methodological Variations

by Duarte Lopes, Isabel Preto and David Freire

Eng. Proc. 2024, 68(1), 27; https://doi.org/10.3390/engproc2024068027 - 9 Jul 2024

Viewed by 338

Abstract

This study investigates optimal training intervals for small hydro power regression models, crucial for accurate forecasts in diverse conditions, particularly focusing on Portugal’s small hydro portfolio. Utilizing a regression model based on kernel density estimation, historical hourly production values, and calendar variables, forecasts [...] Read more.

This study investigates optimal training intervals for small hydro power regression models, crucial for accurate forecasts in diverse conditions, particularly focusing on Portugal’s small hydro portfolio. Utilizing a regression model based on kernel density estimation, historical hourly production values, and calendar variables, forecasts are generated. Various approaches, including dynamic time warping (DTW), “K-Means Alike,” and traditional K-means clustering, are assessed for determining the most effective historical training periods. Results highlight the “K-Means Alike” approach, which, with a 2-month training period, outperforms conventional methods, offering enhanced accuracy while minimizing computational resources. Despite promising results, DTW exhibits increased computational demands without consistent performance superiority. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

6 pages, 344 KiB

Open AccessProceeding Paper

Assessing the Preprocessing Benefits of Data-Driven Decomposition Methods for Phase Permutation Entropy—Application to Econometric Time Series

by Erwan Pierron and Meryem Jabloun

Eng. Proc. 2024, 68(1), 28; https://doi.org/10.3390/engproc2024068028 - 9 Jul 2024

Viewed by 303

Abstract

This paper investigates the efficacy of various data-driven decomposition methods combined with Phase Permutation Entropy (PPE) to form a promising complexity metric for analyzing time series. PPE is a variant of classical permutation entropy (PE), while the examined data-driven decomposition methods include Empirical [...] Read more.

This paper investigates the efficacy of various data-driven decomposition methods combined with Phase Permutation Entropy (PPE) to form a promising complexity metric for analyzing time series. PPE is a variant of classical permutation entropy (PE), while the examined data-driven decomposition methods include Empirical Mode Decomposition (EMD), Variational Mode Decomposition (VMD), Empirical Wavelet Transform (EWT), Seasonal and Trend decomposition using Loess (STL), and Singular Spectrum Analysis-based decomposition (SSA). To our knowledge, this combination has not been explored yet. Our primary aim is to assess how these preprocessing methods affect PPE’s ability to capture temporal structural complexities within time series. This evaluation encompasses the analysis of both simulated and econometric time series. Our results reveal that combining SSA with PPE produces superior advantages for measuring the complexity of seasonal time series. Conversely, VMD combined with PPE proves to be the less advantageous strategy. Overall, our study illustrates that combining data-driven preprocessing methods with PPE offers greater benefits compared to combining them with traditional PE in quantifying time series complexity. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 764 KiB

Open AccessProceeding Paper

Smart Belay Device for Sport Climbing—An Analysis about Falling

by Heiko Oppel and Michael Munz

Eng. Proc. 2024, 68(1), 29; https://doi.org/10.3390/engproc2024068029 - 9 Jul 2024

Viewed by 2667

Abstract

Although sport climbing is generally considered a comparatively safe sport, the injuries that do occur can be more severe than in other sports. With its increased popularity over the last few decades, sport climbing has started to attract people who are unfamiliar with [...] Read more.

Although sport climbing is generally considered a comparatively safe sport, the injuries that do occur can be more severe than in other sports. With its increased popularity over the last few decades, sport climbing has started to attract people who are unfamiliar with the sport. Especially as a belayer, you are responsible for the climber’s safety. For this reason, addressing the issue of safe belaying is necessary. Using an instrumented belay device, we recorded a climber’s fall into the rope with varying configurations. By combining this data with machine learning algorithms, we were able to categorize the recordings and extract information about the severity of the fall for the climber. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Comparison between the two conducted configurations of dynamic and non-dynamic belaying. In case of a fall, the belayer is always pulled in the direction of the rope (first quickdraw), regardless of the type of belaying. The difference is marked by the movement of the belayer as soon as the fall takes place. From there, the belayer must jump in the direction of the rope. This is referred to as dynamic belaying. Without the additional jump, the belayer would be passively standing by, which is referenced as non-dynamic belaying. Full article ">Figure 2
Example of a fall with 0.5 m of slack in the system and a non-dynamic belayer. The timestamp of the impact force (green) and the start time (purple) are marked by vertical lines. Full article ">Figure 3
Visualization of the two fall configurations—slack of 0.5 m (a) and no slack (b). Each subgraph displays a summary of the resulting accelerations, once from dynamic and once from non-dynamic belayed falls. The lines highlighted in the legend represent the average over all of the recordings for each configuration. The slope around the curves represent the standard deviation. Full article ">Figure 4
Feature importance based on the permutation feature importance from the classification results using a support vector machine. Full article ">Figure 5
Visualization of the regression results using a support vector machine and manually extraced features. Subfigure (a) visualizes predicted label over the true label, whereas subfigure (b) shows the predicted label over the absolute deviation from the true label. Full article ">

10 pages, 333 KiB

Open AccessProceeding Paper

Energy Efficiency Evaluation of Frameworks for Algorithms in Time Series Forecasting

by Sergio Aquino-Brítez, Pablo García-Sánchez, Andrés Ortiz and Diego Aquino-Brítez

Eng. Proc. 2024, 68(1), 30; https://doi.org/10.3390/engproc2024068030 - 9 Jul 2024

Viewed by 656

Abstract

In this study, the energy efficiency of time series forecasting algorithms is addressed in a broad context, highlighting the importance of optimizing energy consumption in computational applications. The purpose of this study is to compare the energy efficiency and accuracy of algorithms implemented [...] Read more.

In this study, the energy efficiency of time series forecasting algorithms is addressed in a broad context, highlighting the importance of optimizing energy consumption in computational applications. The purpose of this study is to compare the energy efficiency and accuracy of algorithms implemented in different frameworks, specifically Darts, TensorFlow, and Prophet, using the ARIMA technique. The experiments were conducted on a local infrastructure. The Python library CodeCarbon and the physical energy consumption measurement device openZmeter were used to measure the energy consumption. The results show significant differences in energy consumption and algorithm accuracy depending on the framework and execution environment. We conclude that it is possible to achieve an optimal balance between energy efficiency and accuracy in time series forecasting, which has important implications for developing more sustainable and efficient applications. This study provides valuable guidance for researchers and professionals interested in the energy efficiency of forecasting algorithms. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 636 KiB

Open AccessProceeding Paper

Multimodal Model Based on LSTM for Production Forecasting in Oil Wells with Rod Lift System

by David Esneyder Bello Angulo and Elizabeth León Guzmán

Eng. Proc. 2024, 68(1), 31; https://doi.org/10.3390/engproc2024068031 - 10 Jul 2024

Viewed by 322

Abstract

This paper presents a novel multimodal recurrent model for time series forecasting leveraging LSTM architecture, with a focus on production forecasting in oil wells equipped with rod lift systems. The model is specifically designed to handle time series data with diverse types, incorporating [...] Read more.

This paper presents a novel multimodal recurrent model for time series forecasting leveraging LSTM architecture, with a focus on production forecasting in oil wells equipped with rod lift systems. The model is specifically designed to handle time series data with diverse types, incorporating both images and numerical data at each time step. This capability enables a comprehensive analysis over specified temporal windows. The architecture consists of distinct submodels tailored to process different data modalities. These submodels generate a unified concatenated feature vector, providing a holistic representation of the well’s operational status. This representation is further refined through a dense layer to facilitate non-linear transformation and integration. Temporal analysis forms the core of the model’s functionality, facilitated by a Long Short-Term Memory (LSTM) layer, which excels at capturing long-range dependencies in the data. Additionally, a fully connected layer with linear activation output enables one-shot multi-step forecasting, which is necessary because the input and output have different modalities. Experimental results show that the proposed multimodal model achieved the best performance in the studied cases, with a Mean Absolute Percentage Error (MAPE) of 8.2%, outperforming univariate and multivariate deep learning-based models, as well as ARIMA implementations, which yielded results with a MAPE greater than 9%. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 17163 KiB

Open AccessProceeding Paper

Spectral Characteristics of Strong Ground Motion Time Series for Low to Medium Seismicity Regions with Deep Soil Atop Deep Geological Sediments

by Silva Lozančić, Borko Đ. Bulajić, Gordana Pavić, Ivana Bulajić and Marijana Hadzima-Nyarko

Eng. Proc. 2024, 68(1), 32; https://doi.org/10.3390/engproc2024068032 - 10 Jul 2024

Viewed by 266

Abstract

In this study, we examine the features of horizontal and vertical strong motion time series for sites with deep soil above deep geological deposits, in low-to-medium seismicity zones. New empirical regional equations were derived for spectral and peak ground acceleration attenuation based only [...] Read more.

In this study, we examine the features of horizontal and vertical strong motion time series for sites with deep soil above deep geological deposits, in low-to-medium seismicity zones. New empirical regional equations were derived for spectral and peak ground acceleration attenuation based only on the strong-motion time series that were recorded in the north-western Balkans region. Results show that short-period horizontal spectral amplitudes at rock locations can be larger than those resulting from the combination of the deep soil and the deep geological deposits. However, the results also show that there is a significant increase in spectral amplitudes for larger vibration periods. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Pannonian Basin (top left), epicenters of Mw ≥ 3 regional earthquakes (top right), the investigated area, the epicenters of two recent damaging earthquakes, and the strongest historical earthquake close to Osijek [<a href="#B5-engproc-68-00032" class="html-bibr">5</a>]. Full article ">Figure 2
Two typical geotechnical profiles and the geological map for the analyzed region. Full article ">Figure 3
Regional horizontal pseudo-acceleration spectral records at deep soil sites (sL = 2) and the corresponding empirical predictions based on the new scaling equations for horizontal ground motion [<a href="#B5-engproc-68-00032" class="html-bibr">5</a>]. Full article ">Figure 4
Regional vertical pseudo-acceleration spectral records at deep soil sites (sL = 2) and the corresponding empirical predictions based on the new scaling equations for vertical ground motion [<a href="#B8-engproc-68-00032" class="html-bibr">8</a>]. Full article ">Figure 5
Microzonation maps for 6 horizontal PSA amplitudes and probabilities of p in t years that are analog to return periods of 95 years (top left), 475 years (top right), 975 years (bottom left), and 2475 years (bottom right), for the deep soil and deep geological sediments; the location for which the UHS were calculated is shown by full circles [<a href="#B5-engproc-68-00032" class="html-bibr">5</a>]. Full article ">Figure 6
UHS for the location 45°32′ N, 18°23′ E for four different probabilities [<a href="#B5-engproc-68-00032" class="html-bibr">5</a>], and the Eurocode 8 (2004) spectra for ground type C and Type 2. Full article ">Figure 7
UHS vertical spectra for four distinct probabilities [<a href="#B8-engproc-68-00032" class="html-bibr">8</a>] and the corresponding Eurocode 8 (2004) spectra for Ground Type C and Type 2, for the location 45 32′ N, 18 23′′ E. Full article ">

10 pages, 678 KiB

Open AccessProceeding Paper

Performance of an End-to-End Inventory Demand Forecasting Pipeline Using a Federated Data Ecosystem

by Henrique Duarte Moura, Els de Vleeschauwer, Gerald Haesendonck, Ben De Meester, Lynn D’eer, Tom De Schepper, Siegfried Mercelis and Erik Mannens

Eng. Proc. 2024, 68(1), 33; https://doi.org/10.3390/engproc2024068033 - 10 Jul 2024

Viewed by 356

Abstract

One of the key challenges for (fresh produce) retailers is achieving optimal demand forecasting, as it plays a crucial role in operational decision-making and dampens the Bullwhip Effect. Improved forecasts holds the potential to achieve a balance between minimizing waste and avoiding shortages. [...] Read more.

One of the key challenges for (fresh produce) retailers is achieving optimal demand forecasting, as it plays a crucial role in operational decision-making and dampens the Bullwhip Effect. Improved forecasts holds the potential to achieve a balance between minimizing waste and avoiding shortages. Different retailers have partial views on the same products, which—when combined—can improve the forecasting of individual retailers’ inventory demand. However, retailers are hesitant to share all their individual data. Therefore, we propose an end-to-end graph-based time series forecasting pipeline using a federated data ecosystem to predict inventory demand for supply chain retailers. Graph deep learning forecasting has the ability to comprehend intricate relationships, and it seamlessly tunes into the diverse, multi-retailer data present in a federated setup. The system aims to create a unified data view without centralization, addressing technical and operational challenges, which are discussed throughout the text. We test this pipeline using real-world data across large and small retailers, and discuss the performance obtained and how it can be further improved. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
BWE (red) generates an amplification in the inventory swings alongside the supply chain in response to changes in consumer demand. The purple curve shows how data sharing and enhanced forecasting damps BWE in the whole chain (SOURCE: imec ICON project AI4Foodlogistics proposal). Full article ">Figure 2
RML rules describe how source data are mapped to RDF graphs using a common ontology. RML mapping engines execute the rules. SPARQL endpoints expose the RDF graphs. An additional SPARQL Endpoint (Federator) combines the exposed RDF graphs to feed the forecasting model that generates order predictions. Retailers stay in control of their own data, whilst contributing to a retailer-independent forecasting model. Full article ">Figure 3
Proposed spatio-temporal graph-based deep learning model. Full article ">Figure 4
A time-evolving graph in which four nodes <math display="inline"><semantics> <mrow> <mo>{</mo> <msub> <mi>p</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>p</mi> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi>p</mi> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi>p</mi> <mn>3</mn> </msub> <mo>}</mo> </mrow> </semantics></math> correspond to four products. Given the observations and graph snapshots in the past M time steps, where for each step each node <math display="inline"><semantics> <msub> <mi>p</mi> <mi>i</mi> </msub> </semantics></math> has some feature values, we want to infer in the next <math display="inline"><semantics> <msup> <mi>T</mi> <mo>′</mo> </msup> </semantics></math> time steps the values for these features. The graph can be directed (as above) or undirected (no arrows). Full article ">Figure 5
Forecasting module runtime considering several setups. Full article ">

14 pages, 5554 KiB

Open AccessProceeding Paper

Short-Term Forecasting of Non-Stationary Time Series

by Amir Aieb, Antonio Liotta, Alexander Jacob and Muhammad Azfar Yaqub

Eng. Proc. 2024, 68(1), 34; https://doi.org/10.3390/engproc2024068034 - 10 Jul 2024

Viewed by 450

Abstract

Forecasting climate events is crucial for mitigating and managing risks related to climate change; however, the problem of non-stationarity in time series (NTS) arises, making it difficult to capture and model the underlying trends. This task requires a complex procedure to address the [...] Read more.

Forecasting climate events is crucial for mitigating and managing risks related to climate change; however, the problem of non-stationarity in time series (NTS) arises, making it difficult to capture and model the underlying trends. This task requires a complex procedure to address the challenge of creating a strong model that can effectively handle the non-uniform variability in various climate datasets. In this work, we use a daily standardized precipitation index dataset as an example of NTS, whereby the heterogeneous variability of daily precipitation poses complexities for traditional machine-learning models in predicting future events. To address these challenges, we introduce a novel approach, aiming to adjust the non-uniform distribution and simplify the detection time lags using autocorrelation. Our study employs a range of statistical techniques, including sampling-based seasonality, mathematical transformation, and normalization, to preprocess the data to increase the time lag window. Through the exploration of linear and sinusoidal transformation, we aim to assess their impact on increasing the accuracy of forecasting models. A strong performance is effectively observed by using the proposed approach to capture more than one year of time delay across all the seasonal subsets. Furthermore, improved model accuracy is observed, notably with K-Nearest Neighbors (KNN) and Random Forest (RF). This study underscores RF’s consistently strong performance across all the transformations, while KNN only demonstrates optimal results when the data have been linearized. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Algorithm for transforming non-stationary time series data. Full article ">Figure 2
Statistical description of daily precipitation (A) and standardized precipitation index obtained from the gamma and Pearson_3 models (B), followed by the test of homogeneity for both the training and testing datasets (C,D). Full article ">Figure 3
The flowchart summarizes the workflow of the proposed approach for time series data forecasting. Full article ">Figure 4
Autocorrelation analysis illustrating the time lag detection using various data transformation methods. Full article ">Figure 5
Results of training K-Nearest Neighbors (KNN) and Random Forest (RF) models to forecast the Standardized Precipitation Index (SPI), under different data transformation methods. Full article ">Figure 6
Results of testing the K-Nearest Neighbors (KNN) and Random Forest (RF) models for the standardized precipitation index (SPI) forecasting on the original scale. Full article ">Figure 7
Cross-validation analysis for SPI forecasting via K-Nearest Neighbors (KNN) and Random Forest (RF) using the maximum time delay. Coefficient of determination (R-squared), mean absolute error (MAE). Full article ">

7 pages, 369 KiB

Open AccessProceeding Paper

Modelling Explosive Nonstationarity of Ground Motion Shows Potential for Landslide Early Warning

by Michael Manthey, Guoqi Qian and Antoinette Tordesillas

Eng. Proc. 2024, 68(1), 35; https://doi.org/10.3390/engproc2024068035 - 11 Jul 2024

Viewed by 314

Abstract

This work applies the rarely seen explosive version of autoregressive modelling to a novel practical context—geological failure monitoring. This approach is more general than standard ARMA or ARIMA methods in that it allows the underlying data process to be explosively nonstationary, which is [...] Read more.

This work applies the rarely seen explosive version of autoregressive modelling to a novel practical context—geological failure monitoring. This approach is more general than standard ARMA or ARIMA methods in that it allows the underlying data process to be explosively nonstationary, which is often the case in real-world slope failure processes. We develop and test our methodology on a case study consisting of high-quality (in situ) line-of-sight radar displacement data from a slope that undergoes a failure event. Specifically, we first optimally estimate the characteristic roots of the autoregressive processes underpinning the displacement time series preceding the failure at each monitoring location. We then establish and utilise a pivotal quantity for the autoregressive parameter ensemble to perform simulation-based hypothesis test/s for the explosiveness of the corresponding true characteristic roots. Concluding that a true characteristic root becomes explosive at some significance level implies that the underlying displacement process is explosively nonstationary, and, hence, local geological instability is suspected at this significance level. We found that the actual location of failure (LOF) was identified well in advance of the time of failure (TOF) by flagging those locations where explosive root(s) were identified by our approach. This statistical feedback model for ground motion dynamics presents an alternative and/or addition to the velocity threshold approach to early warning of impending failure. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Mine-A representative cumulative displacement profiles (every 10th location). Note that there are several outlying series with drastic changes in displacement and which do not fall within the LOF. Otherwise, two distinct regimes emerge; a relatively stable grouping of time series and a subset which seem to exponentially diverge. Full article ">Figure 2
Spatial maps of estimated roots over Mine-A locations. (a) Map of Mine-A locations, coloured according to the value of the largest modulus of the characteristic roots estimated at each location. The first 200 training times were used from the total 1315 prefailure times. A clear ovular region of points identified as near-explosive already emerges in the centre-right. (b) Map of Mine-A locations, coloured according to the value of the largest modulus of the characteristic roots estimated at each location. The first 1000 training times were used from the total 1315 prefailure times. Note that it may appear (due to the gradual nature of colouration) to some viewers that a large portion of the slope is now explosive; however, the surrounding regions are typically slightly below a value of one and, hence, are still stationary. In fact, more or less, only those points in the eventual LOF actually sometimes exceed a value of 1. (c) Map of Mine-A locations, coloured according to the value of the largest modulus of the characteristic roots estimated at each location. The first 1310 training times were used from the total 1315 prefailure times. The ovular region of points identified as explosive is now quite stark. Full article ">Figure 3
Spatial maps of stationary versus explosive hypothesis test outcomes over Mine-A locations. (a) Map of Mine-A locations, coloured according to whether our simulation-based hypothesis testing scheme accepts/rejects the null hypothesis of stationarity at a confidence level of 95% (location-wise). The first 200 training times were used from the total 1315 prefailure times. The previously seen ovular region of points identified as explosive is yet to emerge at this high threshold of confidence (however, it does so clearly at lower thresholds of confidence). (b) Map of Mine-A locations, coloured according to whether our simulation-based hypothesis testing scheme accepts/rejects the null hypothesis of stationarity at a confidence level of 95% (location-wise). The first 1000 training times were used from the total 1315 prefailure times. The eventual LOF is clearly emerging. (c) Map of Mine-A locations, coloured according to whether our simulation-based hypothesis testing scheme accepts/rejects the null hypothesis of stationarity at a confidence level of 95% (location-wise). The first 1310 training times were used from the total 1315 prefailure times. The ovular region of points identified as explosive is now clearly identified. Full article ">

8 pages, 679 KiB

Open AccessProceeding Paper

Cellular Automata Framework for Dementia Classification Using Explainable AI

by Siva Manohar Reddy Kesu, Neelam Sinha and Hariharan Ramasangu

Eng. Proc. 2024, 68(1), 36; https://doi.org/10.3390/engproc2024068036 - 11 Jul 2024

Viewed by 346

Abstract

The clinical dementia rating scale has been used for analyzing dementia severity based on cognitive impairments. Many researchers have introduced various statistical methods and machine learning techniques for the classification of dementia severity. Feature importance with deep learning architecture can give a better [...] Read more.

The clinical dementia rating scale has been used for analyzing dementia severity based on cognitive impairments. Many researchers have introduced various statistical methods and machine learning techniques for the classification of dementia severity. Feature importance with deep learning architecture can give a better analysis of the dementia severity. A CA framework has been proposed for the classification of cognitive impairment, and LIME has been used for explaining the local interpretability. Feature vectors for healthy and unhealthy classes have been converted to redistributed CA images. These CA images have been classified using deep learning architecture, and promising results have been achieved. GRAD-CAM and LIME explainer have captured the feature importance of the cognitive impairment of CA images. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
Block diagram of the proposed CA framework-based dementia classification. Full article ">Figure 2
Distribution of features in CA grid. Full article ">Figure 3
Redistributed evolutionary CA image from features with diffusion rate of 20%. Full article ">Figure 4
Block diagram of CNN architecture for CA image classifications. Full article ">Figure 5
Predicted class and GRAD-CAM explainer for a sample CA image. (a) Class predicted as unhealthy for CA image from CNN classifier. Feature 1 (orientation) and feature 2 (judgment) play high importance, represented in yellow, whereas features 0 (memory), feature 3 (community affairs), and feature 4 (hobbies) play low importance, represented in blue. (b) GRAD-CAM image explanation for low cognitive impairment CA image based on the color intensity is high for orientation and judgment features. Full article ">Figure 6
Feature importance from LIME explainer for low cognitive impairment CA image based on the color intensity. The intensity of color is high for orientation, judgment, and community affair features in a decreasing order. Full article ">

14 pages, 390 KiB

Open AccessProceeding Paper

JET: Fast Estimation of Hierarchical Time Series Clustering

by Phillip Wenig, Mathias Höfgen and Thorsten Papenbrock

Eng. Proc. 2024, 68(1), 37; https://doi.org/10.3390/engproc2024068037 - 11 Jul 2024

Viewed by 373

Abstract

Clustering is an effective, unsupervised classification approach for time series analysis applications that suffer a natural lack of training data. One such application is the development of jet engines, which involves numerous test runs and failure detection processes. While effective data mining algorithms [...] Read more.

Clustering is an effective, unsupervised classification approach for time series analysis applications that suffer a natural lack of training data. One such application is the development of jet engines, which involves numerous test runs and failure detection processes. While effective data mining algorithms exist for the detection of anomalous and structurally conspicuous test recordings, these algorithms do not perform any semantic labeling. So, data analysts spend many hours connecting the large amounts of automatically extracted observations to their underlying root causes. The complexity, number, and variety of extracted time series make this task hard not only for humans, but also for existing time series clustering algorithms. These algorithms either require training data for supervised learning, cannot deal with varying time series lengths, or suffer from exceptionally long runtimes. In this paper, we propose JET, an unsupervised, highly efficient clustering algorithm for large numbers of variable-lengths time series. The main idea is to transform the input time series into a metric space, then apply a very fast conventional clustering algorithm to obtain effective but rather coarse-grained pre-clustering of the data; this pre-clustering serves to subsequently estimate the more accurate but also more costly shape-based distances of the time series and, thus, enables JET to apply a highly effective hierarchical clustering algorithm to the entire input time series collection. Our experiments demonstrate that JET is highly accurate and much faster than its competitors. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
The four parts of JET: feature embedding, pre-clustering, distance matrix estimation, and hierarchical clustering. Full article ">Figure 2
Birch’s homogeneity score increases with an increasing number of clusters on a fixed sample dataset. Full article ">Figure 3
Distance calculations for the exact HC and approximating JET. Full article ">Figure 4
Rand Index (RI) scores of all algorithms on the UCR Time Series Classification Archive datasets grouped by datasets of varying lengths (left) and equal lengths (right) and ordered by median RI (on varying lengths). Full article ">Figure 5
(Left): Runtime in seconds of well-performing algorithms (mean RI <math display="inline"><semantics> <mrow> <mo>></mo> <mn>0.5</mn> </mrow> </semantics></math>) on the UCR varying-length datasets ordered by median RI. (Right): Average runtime in seconds of JET and the standard HC algorithm on the UCR datasets with varying lengths plotted against numbers of time series per dataset. Full article ">Figure 6
Rand Index (RI) scores vs. runtime measurements in seconds (on log scale) for two real-world Rolls-Royce datasets. Full article ">

11 pages, 5561 KiB

Open AccessProceeding Paper

A System for Efficient Detection of Forest Fires through Low Power Environmental Data Monitoring and AI

by İpek Üremek, Paul Leahy and Emanuel Popovici

Eng. Proc. 2024, 68(1), 38; https://doi.org/10.3390/engproc2024068038 - 11 Jul 2024

Viewed by 420

Abstract

This study introduces a system that merges AI with low-power IoT (Internet Of Things) technology to enhance environmental monitoring, with a specific focus on accurately predicting forest fires through time series analysis. Utilizing affordable sensors and wireless communication technologies like LoRa (Long Range), [...] Read more.

This study introduces a system that merges AI with low-power IoT (Internet Of Things) technology to enhance environmental monitoring, with a specific focus on accurately predicting forest fires through time series analysis. Utilizing affordable sensors and wireless communication technologies like LoRa (Long Range), environmental data have been gathered. One of the key features of this approach is the comparison of the real-time local environmental data with meteorological service environmental data to ensure accuracy. This comparison informs a feedback loop that improves the model’s predictive accuracy. The research also delves into detailed time series analysis, incorporating the Autoregressive Integrated Moving Average (ARIMA) model to identify the best windows of opportunity for communication and to provide future forecasting. Finally, a decision tree model serves as the last step, providing a comprehensive assessment of fire risk due to its straightforward application and clarity. Validation of the fire detection component remains a critical future task to confirm its effectiveness and reliability. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

5 pages, 887 KiB

Open AccessProceeding Paper

New Studies on Birth, Death, Temperature Time Series and Their Correlation

by Arzu Sardarli

Eng. Proc. 2024, 68(1), 39; https://doi.org/10.3390/engproc2024068039 - 12 Jul 2024

Viewed by 263

Abstract

This article presents the preliminary results of my daily birth time series studies in eight Canadian provinces for 2011–2019. An extensive review of the available literature shows that researchers usually analyse the monthly birth time series, perhaps because of the unavailability of daily [...] Read more.

This article presents the preliminary results of my daily birth time series studies in eight Canadian provinces for 2011–2019. An extensive review of the available literature shows that researchers usually analyse the monthly birth time series, perhaps because of the unavailability of daily birth data. Statistics Canada started releasing daily birth data records for Canadian provinces and territories in 2011. Using the Fourier analysis of daily birth time series, I revealed new cycles on birth time series that, to my knowledge, were not observed in previous works. This research was conducted at the Regina and Saskatoon Research Data Centres (RDC) of Statistics Canada. It was supported by the University of Regina, the First Nations University of Canada and Statistics Canada. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

5 pages, 985 KiB

Open AccessProceeding Paper

Crisis and Youth Inactivity: Central and Eastern Europe during the Financial Crisis of 2008 and the COVID-19 Outbreak of 2020

by Nataša Kurnoga, Tomislav Korotaj and James Ming Chen

Eng. Proc. 2024, 68(1), 40; https://doi.org/10.3390/engproc2024068040 - 12 Jul 2024

Viewed by 425

Abstract

This paper analyzes eleven Central and Eastern European countries after the financial crisis of 2008 and the COVID-19 pandemic of 2020. It investigates the heterogeneity in the labor market among the selected countries based on the youth inactivity, secondary education attainment, and income [...] Read more.

This paper analyzes eleven Central and Eastern European countries after the financial crisis of 2008 and the COVID-19 pandemic of 2020. It investigates the heterogeneity in the labor market among the selected countries based on the youth inactivity, secondary education attainment, and income share of the bottom fifty percent of the population. A hierarchical cluster analysis with Ward’s method and k-means clustering generated diverse cluster solutions. A comparative analysis of the four-cluster solutions for 2008 and 2020 showed multiple changes in the cluster composition. The joint groupings of geographically and historically close countries, such as the Baltics, the former Czechoslovakia, and the former Yugoslav republics of Croatia and Slovenia, were identified for 2008. Lithuania emerged as a singleton in 2020. The youth inactivity, educational levels, and income inequality reveal the status of the youth in Central and Eastern Europe during these crises. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 1876 KiB

Open AccessProceeding Paper

Application of the Optimised Pulse Width Modulation (PWM) Based Encoding-Decoding Algorithm for Forecasting with Spiking Neural Networks (SNNs)

by Sergio Lucas and Eva Portillo

Eng. Proc. 2024, 68(1), 41; https://doi.org/10.3390/engproc2024068041 - 12 Jul 2024

Viewed by 261

Abstract

Spiking Neural Networks (SNNs) are recognised for processing spatiotemporal information with ultra-low power consumption. However, applying a non-efficient encoding-decoding algorithm can counter the efficiency advantages of the SNNs. In this sense, this paper presents one-step ahead forecasting centered on the application of an [...] Read more.

Spiking Neural Networks (SNNs) are recognised for processing spatiotemporal information with ultra-low power consumption. However, applying a non-efficient encoding-decoding algorithm can counter the efficiency advantages of the SNNs. In this sense, this paper presents one-step ahead forecasting centered on the application of an optimised encoding-decoding algorithm based on Pulse Width Modulation (PWM) for SNNs. The validation is carried out with sine-wave, 3 UCI and 1 available real-world datasets. The results show the practical disappearance of the computational and energy costs associated with the encoding and decoding phases (less than 2% of the total costs) and very satisfactory forecasting results (MAE lower than 0.0357) for any dataset. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 325 KiB

Open AccessProceeding Paper

New Algorithm for Detecting Weak Changes in the Mean in a Class of CHARN Models with Application to Welding Electrical Signals

by Youssef Salman, Anis Hoayek and Mireille Batton-Hubert

Eng. Proc. 2024, 68(1), 42; https://doi.org/10.3390/engproc2024068042 - 12 Jul 2024

Viewed by 264

Abstract

In this paper, we propose a new automatic algorithm for detecting weak changes in the mean of a class of piece-wise CHARN models. Through a simulation experiment, we demonstrate its efficacy and precision in detecting weak changes in the mean and accurately estimating [...] Read more.

In this paper, we propose a new automatic algorithm for detecting weak changes in the mean of a class of piece-wise CHARN models. Through a simulation experiment, we demonstrate its efficacy and precision in detecting weak changes in the mean and accurately estimating their locations. Furthermore, we illustrate the robust performance of our algorithm through its application to welding electrical signals (WES). Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 255 KiB

Open AccessProceeding Paper

Detecting Short-Notice Cancellation in Hotels with Machine Learning

by Eleazar C-Sánchez and Agustín J. Sánchez-Medina

Eng. Proc. 2024, 68(1), 43; https://doi.org/10.3390/engproc2024068043 - 15 Jul 2024

Viewed by 322

Abstract

Cancellations play a critical role in the lodging industry. Considering the time horizon, cancellations placed close to check-in have a significant impact on hoteliers, who must respond promptly for effective management. In recent years, the introduction of personal name records (PNR) has brought [...] Read more.

Cancellations play a critical role in the lodging industry. Considering the time horizon, cancellations placed close to check-in have a significant impact on hoteliers, who must respond promptly for effective management. In recent years, the introduction of personal name records (PNR) has brought innovative approaches to this domain, but short-notice cancellation prediction is still underdeveloped. Using real PNR data with more than 10k reservations provided by a four-star hotel, this research aims to combine fuzzy clustering with tree decision techniques and random forest under R software version 4.3.3 to forecast cancellations placed close to the entry day, slightly improving the performance of individual techniques. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

11 pages, 576 KiB

Open AccessProceeding Paper

Shadows of Resilience: Exploring the Impact of the Shadow Economy on Economic Stability

by Charalampos Agiropoulos, James Ming Chen, Thomas Poufinas and George Galanos

Eng. Proc. 2024, 68(1), 44; https://doi.org/10.3390/engproc2024068044 - 15 Jul 2024

Viewed by 389

Abstract

This study analyzes the shadow economy within the European Union and its influence on the economic resilience of member countries. Data spanning almost two decades and covering a broad spectrum provide a unique opportunity to examine the impact of the shadow economy on [...] Read more.

This study analyzes the shadow economy within the European Union and its influence on the economic resilience of member countries. Data spanning almost two decades and covering a broad spectrum provide a unique opportunity to examine the impact of the shadow economy on economic stability across various economic cycles. Regularization techniques such as Lasso and Automatic Relevance Determination (ARD) combat possible collinearity and overfitting arising from the inclusion of irrelevant and redundant variables. The shadow economy interacts with key indicators of economic resilience, such as GDP, national debt, and population, across different phases of economic stability and turbulence. The preliminary findings suggest a complex and varied interaction between the shadow economy and economic resilience. This study provides a valuable foundation for policies aimed at stability and sustainable economic growth. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 619 KiB

Open AccessProceeding Paper

Robust Multitaper Tests for Detecting Frequency Modulated Signals

by Benjamin Ott, Glen Takahara and Wesley S. Burr

Eng. Proc. 2024, 68(1), 45; https://doi.org/10.3390/engproc2024068045 - 15 Jul 2024

Viewed by 323

Abstract

In this paper, we propose a new test for the detection of polynomial modulated signals, as well as an aggregated test based on the new test. The test statistics are developed under the multitaper spectral framework, and are designed to be robust to [...] Read more.

In this paper, we propose a new test for the detection of polynomial modulated signals, as well as an aggregated test based on the new test. The test statistics are developed under the multitaper spectral framework, and are designed to be robust to the choice of the multitaper order K. The proposed tests are based on a modification of a previous test statistic developed by the authors, denoted by

F_{4}

. We review the

F_{4}

test, and discuss some of its shortcomings, then propose our two tests. We illustrate performance via simulations, and apply our tests to SoHO GOLF optical solar time series. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 2121 KiB

Open AccessProceeding Paper

Optimizing Biogas Power Plants through Machine-Learning-Aided Rotor Configuration

by Andreas Heller, Héctor Pomares and Peter Glösekötter

Eng. Proc. 2024, 68(1), 46; https://doi.org/10.3390/engproc2024068046 - 16 Jul 2024

Viewed by 331

Abstract

The increasing demand for sustainable energy sources has intensified the exploration of biogas power plants as a viable option. In this research, we present a novel approach that leverages machine learning techniques to optimize the performance of biogas power plants through the strategic [...] Read more.

The increasing demand for sustainable energy sources has intensified the exploration of biogas power plants as a viable option. In this research, we present a novel approach that leverages machine learning techniques to optimize the performance of biogas power plants through the strategic placement and configuration of rotors within the fermentation vessel. Our study involves the simulation of a diverse range of biogas power plant scenarios, each characterized by varying rotor locations and rotating speeds, influencing the agitation levels of the biogas substrate. The simulation results, encompassing multiple performance metrics, serve as input data for an artificial neural network (ANN). This ANN is trained to learn the intricate relationships between rotor placement, rotor speed, agitation levels, and overall system efficiency. The trained model demonstrates predictive capabilities, enabling the estimation of plant efficiency based on specific rotor configurations. The proposed methodology provides a tool for both optimizing existing biogas power plants and guiding engineers in the design and setup of new facilities. Our model aims to offer valuable insights for engineers in the initial planning stages of new biogas power plants, enabling them to make informed decisions that contribute to sustainable and efficient energy generation. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 337 KiB

Open AccessProceeding Paper

A Hybrid Computer-Intensive Approach Integrating Machine Learning and Statistical Methods for Fake News Detection

by Livio Fenga

Eng. Proc. 2024, 68(1), 47; https://doi.org/10.3390/engproc2024068047 - 16 Jul 2024

Viewed by 337

Abstract

In this paper, we address the challenge of early fake news detection within the framework of anomaly detection for time-dependent data. Our proposed method is computationally intensive, leveraging a resampling scheme inspired by maximum entropy principles. It has a hybrid nature, combining a [...] Read more.

In this paper, we address the challenge of early fake news detection within the framework of anomaly detection for time-dependent data. Our proposed method is computationally intensive, leveraging a resampling scheme inspired by maximum entropy principles. It has a hybrid nature, combining a sophisticated machine learning algorithm augmented by a bootstrapped versions of binomial statistical tests. In the presented approach, the detection of fake news through the anomaly detection system entails identifying sudden deviations from the norm, indicative of significant, temporary shifts in the underlying data-generating process. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 337 KiB

Open AccessProceeding Paper

Multi-Objective Optimisation for the Selection of Clusterings across Time

by Sergej Korlakov, Gerhard Klassen, Luca T. Bauer and Stefan Conrad

Eng. Proc. 2024, 68(1), 48; https://doi.org/10.3390/engproc2024068048 - 17 Jul 2024

Viewed by 289

Abstract

Nowadays, time series data are ubiquitous, encompassing various domains like medicine, economics, energy, climate science and the Internet of Things. One crucial task in analysing these data is clustering, aiming to find patterns that indicate previously undiscovered relationships among features or specific groups [...] Read more.

Nowadays, time series data are ubiquitous, encompassing various domains like medicine, economics, energy, climate science and the Internet of Things. One crucial task in analysing these data is clustering, aiming to find patterns that indicate previously undiscovered relationships among features or specific groups of objects. In this work, we present a novel framework for the clustering of multiple multivariate time series over time that utilises multi-objective optimisation to determine the temporal clustering solution for each time point. To highlight the strength of our framework, we conduct a comparison with alternative solutions using multiple labelled real-world datasets. Our results reveal that our method not only provides better results but also enables a comparison between datasets with regard to their temporal dependencies. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

10 pages, 1062 KiB

Open AccessProceeding Paper

Continual Learning for Time Series Forecasting: A First Survey

by Quentin Besnard and Nicolas Ragot

Eng. Proc. 2024, 68(1), 49; https://doi.org/10.3390/engproc2024068049 - 17 Jul 2024

Viewed by 495

Abstract

Deep learning has brought significant advancements in the field of artificial intelligence, particularly in robotics, imaging, sound processing, etc. However, a common major challenge faced by all neural networks is their substantial demand for data during the learning process. The required data must [...] Read more.

Deep learning has brought significant advancements in the field of artificial intelligence, particularly in robotics, imaging, sound processing, etc. However, a common major challenge faced by all neural networks is their substantial demand for data during the learning process. The required data must be both quantitative and stationary to ensure the proper computing of standard models. Nevertheless, complying to these constraints is often impossible for many real-life applications because of dynamic environments. Indeed, modifications can occur in the distribution of the data or even in the goals to pursue within these environments. This is known as data and concept drift. Research in the field of continual learning seeks to address these challenges by implementing evolving models capable of adaptation over time. This notably involves finding a compromise on the plasticity/stability dilemma while taking into account material and computational constraints. Exploratory efforts are evident in all applications of deep learning (graphs, reinforcement learning, etc.), but to date, there is still a limited amount of work in the case of time series, specifically in the context of regression and forecasting. This paper aims to provide a first survey on this field of continuous learning applied to time series forecasting. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

17 pages, 851 KiB

Open AccessProceeding Paper

A Machine Learning-Based Approach to Analyze and Visualize Time-Series Sentencing Data

by Eugene Pinsky and Kandaswamy Piranavakumar

Eng. Proc. 2024, 68(1), 50; https://doi.org/10.3390/engproc2024068050 - 17 Jul 2024

Viewed by 240

Abstract

Analyzing time-series sentencing data presents many challenges. The data have many dimensions and change with time. This makes it difficult to identify patterns and discuss their similarities over time. This work proposes a machine learning approach to associate patterns with clusters. This allows [...] Read more.

Analyzing time-series sentencing data presents many challenges. The data have many dimensions and change with time. This makes it difficult to identify patterns and discuss their similarities over time. This work proposes a machine learning approach to associate patterns with clusters. This allows a representation of sentencing data regarding trajectories in the appropriate (time, cluster) space. We propose to use the Hamming distance of trajectories to measure the similarity of sentencing data across districts. For any offense, we can define the average Hamming distance that has a simple interpretation as the average period when sentencing patterns are different. We introduce simple statistical measures on trajectories to show similarities and changes in sentencing behavior over time. We illustrate our approach by analyzing sentencing data for narcotics and retail theft. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

12 pages, 1446 KiB

Open AccessProceeding Paper

Assessing Global Wildfire Dynamics and Climate Resilience: A Focus on European Regions Using the Fire Weather Index

by Ayat-Allah Bouramdane

Eng. Proc. 2024, 68(1), 51; https://doi.org/10.3390/engproc2024068051 - 18 Jul 2024

Viewed by 382

Abstract

Wildfires pose significant threats to ecosystems, human safety, and socio-economic stability, necessitating a deep understanding of fire-prone landscapes for effective management. This study assesses the temporal and spatial patterns of the Fire Weather Index (FWI), a crucial indicator of landscape flammability, with a [...] Read more.

Wildfires pose significant threats to ecosystems, human safety, and socio-economic stability, necessitating a deep understanding of fire-prone landscapes for effective management. This study assesses the temporal and spatial patterns of the Fire Weather Index (FWI), a crucial indicator of landscape flammability, with a particular focus on European regions. Historical FWI data from the European Forest Fire Information System (EFFIS) under the Copernicus Emergency Management Service (CEMS) are analyzed using tools such as the Climate Data Store (CDS) API. The results reveal spatial patterns, highlighting regions with heightened wildfire risk and those with reduced fire danger. Southern and Southeastern Europe face elevated danger, driven by factors like high temperatures, low humidity, and reduced precipitation, while Northwestern and Northeastern Europe exhibit lower risk due to milder conditions. The study further delves into the implications of these patterns on agrivoltaic systems, the distinct climatic and environmental factors influencing elevated FWI levels across various regions, and how the findings of this research can guide tailored wildfire management strategies for European areas. The findings inform resilient strategies for policymakers, land managers, and communities, contributing valuable insights for proactive and sustainable wildfire mitigation. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
The global analysis of the Fire Weather Index (FWI) anomaly plot for 2022 reveals distinct patterns indicating regions with increased wildfire risk (i.e., >50, highlighted in red) and those experiencing a relative reduction in fire danger (highlighted in blue). Elevated fire risk is observed across continents, including parts of North and South America, Africa, Asia, and regions with arid climates, suggesting heightened susceptibility to wildfires due to factors such as high temperatures and reduced humidity. Conversely, temperate zones and areas with consistent precipitation, like Northern Europe and coastal South America, exhibit a relative reduction in fire danger. These patterns underscore the complex interplay between climatic conditions and vegetation characteristics, emphasizing the need for tailored wildfire management strategies on a global scale to address the diverse challenges associated with fire risk and resilience. Source: own elaboration based on <a href="#sec3-engproc-68-00051" class="html-sec">Section 3</a>. Full article ">Figure 2
The analysis of the Fire Weather Index (FWI) anomaly over Europe (Top-Left Panel) for the year 2022 unveils distinctive patterns indicating regions with heightened wildfire risk (highlighted in red) and those experiencing a relative reduction in fire danger (highlighted in blue). In Southern and Southeastern Europe (Top-Right Panel), the notable red regions suggest an increased number of days with FWI surpassing the threshold of 50, attributed to warmer temperatures, lower humidity, and decreased precipitation during the summer months. Complex topography, such as mountainous terrain, amplifies fire danger in these areas. Conversely, notable blue regions in Northwestern (Bottom-Left Panel) and Northeastern (Bottom-Right Panel) Europe indicate a reduction in FWI, potentially linked to more moderate temperatures, higher humidity, or increased precipitation. Regions with ample vegetation moisture content and climate resilience, particularly those with consistent precipitation, show reduced fire danger. This nuanced analysis underscores the importance of tailoring wildfire management strategies to the specific climatic and geographical characteristics of different European regions, supporting targeted interventions for effective risk mitigation. Source: own elaboration based on <a href="#sec3-engproc-68-00051" class="html-sec">Section 3</a>. Full article ">Figure 3
The weekly time series plots comparing Fire Weather Index (FWI) anomalies in 2022 with the reference period mean have provided valuable insights into the temporal patterns of fire danger across various European regions—Europe, Southwestern Europe, Northwestern Europe, Southeastern Europe, and Northeastern Europe. The x-axis represents weeks, showcasing mean FWI values, while shaded regions indicate the reference period’s FWI variability. A solid line denotes the reference period mean, and shaded areas above and below represent the 10th to 90th percentile range and the minimum to maximum range during the reference period. These analyses identified specific weeks and regions susceptible to heightened fire risk and those exhibiting a reduced risk. Regions highlighted in red on the plots indicated an excess of FWI values in 2022 compared to the reference period mean, signifying heightened fire danger during those specific weeks. On the other hand, regions depicted in blue exhibited FWI values below the reference period mean, suggesting a relative reduction in fire danger during those weeks. The dynamic perspective offered by these weekly analyses enhances our ability to pinpoint temporal variations in fire risk, supporting proactive measures and emphasizing the importance of region-specific wildfire management strategies. Source: own elaboration based on <a href="#sec3-engproc-68-00051" class="html-sec">Section 3</a>. Full article ">

9 pages, 781 KiB

Open AccessProceeding Paper

Assessing Asymptotic Tail Independence: A Simulation Study

by Marta Ferreira

Eng. Proc. 2024, 68(1), 52; https://doi.org/10.3390/engproc2024068052 - 18 Jul 2024

Viewed by 321

Abstract

The occurrence of extreme values in one variable can trigger the same in other variables, making it necessary to assess the risk of contagion. The usual dependence measures based on the central part of the data typically fail to assess extreme dependence. Within [...] Read more.

The occurrence of extreme values in one variable can trigger the same in other variables, making it necessary to assess the risk of contagion. The usual dependence measures based on the central part of the data typically fail to assess extreme dependence. Within the scope of EVT, tail dependence measures were developed, such as the Ledford and Tawn coefficient that we discuss here. This is a measure of residual dependence that is particularly important when it comes to analyzing at the tail level where data are scarce. We will consider different estimation methodologies and compare them based on a simulation study. We finish with an application to real data. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

15 pages, 3374 KiB

Open AccessProceeding Paper

Exploring Regional Determinants of Tourism Success in the Eurozone: An Unsupervised Machine Learning Approach

by Charalampos Agiropoulos, James Ming Chen, George Galanos and Thomas Poufinas

Eng. Proc. 2024, 68(1), 53; https://doi.org/10.3390/engproc2024068053 - 19 Jul 2024

Viewed by 256

Abstract

This paper presents an initial analysis of the factors influencing tourism success at the NUTS 2 regional level across the Eurozone from 2010 to 2019. Utilizing an extensive dataset that includes economic, demographic, and tourism-specific indicators, we employ unsupervised machine learning techniques, primarily [...] Read more.

This paper presents an initial analysis of the factors influencing tourism success at the NUTS 2 regional level across the Eurozone from 2010 to 2019. Utilizing an extensive dataset that includes economic, demographic, and tourism-specific indicators, we employ unsupervised machine learning techniques, primarily K-means clustering and Principal Component Analysis (PCA), to unearth underlying patterns and relationships. Our study reveals distinct clusters of regions characterized by varying degrees of economic prosperity, infrastructure development, and tourism activity. Through K-means clustering, we identified optimal groupings of regions that share similar characteristics in terms of GDP per capita, unemployment rates, tourist arrivals, and overnight stays, among other metrics. Subsequent PCA provided deeper insights into the most influential factors driving these clusters, offering a reduced-dimensional perspective that highlights the primary axes of variation. The findings underscore significant disparities in tourism success across the Eurozone, with economic robustness and strategic infrastructural investments emerging as key drivers. Regions with higher GDP per capita and lower unemployment rates tend to exhibit higher tourism metrics, suggesting that economic health is a substantial contributor to regional tourism appeal and capacity. This paper contributes to the literature by demonstrating how machine learning can be applied to regional tourism data to better understand and strategize for tourism development. The insights garnered from this study are poised to assist policy-makers and tourism planners in crafting targeted interventions aimed at enhancing tourism competitiveness in underperforming regions. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

7 pages, 1892 KiB

Open AccessProceeding Paper

Simulating the Aerial Ballet: The Dance of Fire-Fighting Planes and Helicopters

by Juha Alander, Lauri Honkasilta and Kalle Saastamoinen

Eng. Proc. 2024, 68(1), 54; https://doi.org/10.3390/engproc2024068054 - 19 Jul 2024

Viewed by 335

Abstract

This study introduces a simulation model to analyze the efficacy of different aerial firefighting strategies in Finland, focusing on the comparative water production capacity and associated costs of firefighting aircraft versus helicopters of varying sizes. By utilizing publicly available data and direct inquiries, [...] Read more.

This study introduces a simulation model to analyze the efficacy of different aerial firefighting strategies in Finland, focusing on the comparative water production capacity and associated costs of firefighting aircraft versus helicopters of varying sizes. By utilizing publicly available data and direct inquiries, the model evaluates the impact of water collection distance on the volume of extinguishing water procured and its costs. The simulation reveals that firefighting aircraft offer a cost-effective solution, particularly when collecting water from distances of thirteen kilometers, where their cost per liter of water aligns with that of smaller helicopters operating closer to the fire zone. The study underscores the importance of precise data input into the calculator, highlighting the potential of aerial firefighting strategies in enhancing wildfire suppression efforts. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 1400 KiB

Open AccessProceeding Paper

Evaluation of Economic Interventions in Economic Blocks during an Economic and Sanitary Crisis

by Carmin Montante and Clemente Hernandez-Rodriguez

Eng. Proc. 2024, 68(1), 55; https://doi.org/10.3390/engproc2024068055 - 19 Jul 2024

Viewed by 281

Abstract

The purpose of this study is to evaluate the economic interventions that took place during the initial stages of the pandemic in 2020 in the US, Mexico, and Canada. These countries share a free trade agreement that indicates their willingness to cooperate in [...] Read more.

The purpose of this study is to evaluate the economic interventions that took place during the initial stages of the pandemic in 2020 in the US, Mexico, and Canada. These countries share a free trade agreement that indicates their willingness to cooperate in economic terms with each other and that they should adopt similar economic policies due to both their shared agreements and proximity. However, the economic interventions adopted by two of the three countries were not considered by the other, which makes for an interesting comparison. Interrupted time series analysis is a quasi-quantitative method that has recently been used in evaluating policy during a specific time. This study is interested in focusing on the economic interventions that were put into practice in neighboring countries that have formed a free trade alliance named USMCA. The method of a systematic analysis of interrupted time series will be used as a basis for organizing the article to provide further validity to the study. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 306 KiB

Open AccessProceeding Paper

Modeling the Future of Hydroelectric Power: A Cross-Country Study

by Farooq Ahmad, Livio Finos and Mariangela Guidolin

Eng. Proc. 2024, 68(1), 56; https://doi.org/10.3390/engproc2024068056 - 19 Jul 2024

Viewed by 460

Abstract

This paper examines the role of hydropower in the context of the energy transition, using innovation diffusion models. The study analyzes time series data of hydropower generation from 1965 to 2022 by applying diffusion models and some other models, such as Prophet and [...] Read more.

This paper examines the role of hydropower in the context of the energy transition, using innovation diffusion models. The study analyzes time series data of hydropower generation from 1965 to 2022 by applying diffusion models and some other models, such as Prophet and ARIMA, for comparison purposes. The models are evaluated across diverse geographic regions, including America, Africa, Europe, Asia, and the Middle East, to determine their effectiveness in predicting hydropower generation trends. The analysis reveals that the GGM consistently outperforms other models in accuracy across all regions. In most cases, the GGM exhibits better performance compared to the Bass, ARIMA, and Prophet models, highlighting its potential as a robust forecasting tool for hydropower generation. This study emphasizes the critical role of accurate forecasting in energy planning and calls for further research to validate these findings and explore additional factors influencing hydropower generation evolution. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 696 KiB

Open AccessProceeding Paper

Catalyzing Supply Chain Evolution: A Comprehensive Examination of Artificial Intelligence Integration in Supply Chain Management

by Sarthak Pattnaik, Natasya Liew, Ali Ozcan Kures, Eugene Pinsky and Kathleen Park

Eng. Proc. 2024, 68(1), 57; https://doi.org/10.3390/engproc2024068057 - 22 Jul 2024

Viewed by 525

Abstract

The integration of Artificial Intelligence (AI) into Supply-Chain Management (SCM) has revolutionized operations, offering avenues for enhanced efficiency and decision-making. AI has become pivotal in tackling various Supply-Chain Management challenges, notably enhancing demand forecasting precision and automating warehouse operations for improved efficiency and [...] Read more.

The integration of Artificial Intelligence (AI) into Supply-Chain Management (SCM) has revolutionized operations, offering avenues for enhanced efficiency and decision-making. AI has become pivotal in tackling various Supply-Chain Management challenges, notably enhancing demand forecasting precision and automating warehouse operations for improved efficiency and error reduction. However, a critical debate arises concerning the choice between less accurate explainable models and more accurate yet unexplainable models in Supply-Chain Management applications. This paper explores this debate within the context of various Supply-Chain Management challenges and proposes a methodology for developing models tailored to different Supply-Chain Management problems. Drawing from academic research and modelling, the paper discusses the applications of AI in demand forecasting, inventory optimization, warehouse automation, transportation management, supply chain planning, supplier management, quality control, risk management, and customer service. Additionally, it examines the trade-offs between model interpretability and accuracy, highlighting the need for a nuanced approach. The proposed methodology advocates for the development of explainable models for tasks where interpretability is crucial, such as risk management and supplier selection, while leveraging unexplainable models for tasks prioritizing accuracy, like demand forecasting and predictive maintenance. Through this approach, stakeholders gain insights into Supply-Chain Management processes, fostering better decision-making and accountability. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

8 pages, 9556 KiB

Open AccessProceeding Paper

Calibration-Free Current Measurement with Integrated Quantum Sensor

by Jens Pogorzelski, Ludwig Horsthemke, Jonas Homrighausen, Dennis Stiegekötter, Frederik Hoffmann, Ann-Sophie Bülter, Markus Gregor and Peter Glösekötter

Eng. Proc. 2024, 68(1), 58; https://doi.org/10.3390/engproc2024068058 - 22 Jul 2024

Viewed by 524

Abstract

This paper presents the application of a compact and fully integrated LED quantum sensor based on the NV centers in diamond for current measurement in a busbar. The magnetic field measurements from the sensor are directly compared with measurements from a numerical simulation, [...] Read more.

This paper presents the application of a compact and fully integrated LED quantum sensor based on the NV centers in diamond for current measurement in a busbar. The magnetic field measurements from the sensor are directly compared with measurements from a numerical simulation, eliminating the need for calibration. The sensor setup achieves an accuracy of 0.28% in the measurement range of 0–30

A

DC. The integration of advanced quantum sensing technology with practical current measurement demonstrates the potential of this sensor for applications in electrical and distribution networks. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

14 pages, 1233 KiB

Open AccessProceeding Paper

A Simple Computational Approach to Predict Long-Term Hourly Electric Consumption

by Eugene Pinsky, Etienne Meunier, Pierre Moreau and Tanvi Sharma

Eng. Proc. 2024, 68(1), 59; https://doi.org/10.3390/engproc2024068059 - 23 Jul 2024

Viewed by 296

Abstract

By exploiting the patterns in past data points, we could forecast long-term consumption with a computationally simple algorithm. Our approach is simple to interpret. It incorporates the seasonality of past consumption and can predict power consumption for any time scale. The algorithm can [...] Read more.

By exploiting the patterns in past data points, we could forecast long-term consumption with a computationally simple algorithm. Our approach is simple to interpret. It incorporates the seasonality of past consumption and can predict power consumption for any time scale. The algorithm can be easily implemented directly in SQL. It can run sub-second long-term predictions on large-scale data marts. The proposed method scored a Mean Absolute Percentage Error (MAPE) of just 5.88% when predicting hourly values for France’s electric consumption in 2017 based on hourly data from 2008 to 2011. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 1511 KiB

Open AccessProceeding Paper

Promoting Electric Vehicle Growth through Infrastructure and Policy: A Forecasting Analysis

by Anuva Banwasi, Adele M. Sinai and Brennan Xavier McManus

Eng. Proc. 2024, 68(1), 60; https://doi.org/10.3390/engproc2024068060 - 18 Jul 2024

Viewed by 226

Abstract

This study examines electric vehicle (EV) adoption in the United States, specifically the interconnected relationship between EV-promoting policies, EV charging infrastructure, and registrations of EVs. Gasoline-powered vehicles make up a significant portion of the US’s carbon emissions, and increasing the use of EVs [...] Read more.

This study examines electric vehicle (EV) adoption in the United States, specifically the interconnected relationship between EV-promoting policies, EV charging infrastructure, and registrations of EVs. Gasoline-powered vehicles make up a significant portion of the US’s carbon emissions, and increasing the use of EVs is a way to decrease this footprint. Over the past decade, there have been many incentives and policy-driven changes to propel electric vehicle adoption forward. The focus of this study is to identify if there a significant relationship between these three factors and the extent to which these factors are significant predictors for each other. To do so, we conduct several statistical tests to analyze the forecasting effect of changes in EV policies on EV infrastructure, and changes in infrastructure on EV registrations. We find that there are significant forecasting relationships between these factors. Furthermore, it is possible to accurately forecast changes in EV charging stations over time using the time-series data of previous EV charging stations and policies. There are many interconnected factors, but this strong forecasting relationship between EV incentive policies and the expansion of charging infrastructure provides valuable insights for policymakers, industry stakeholders, and researchers attempting to understand and promote EV adoption. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Figure 1
(a) Number of EV charging stations in New York (NY) over time; (b) number of EV-related incentive policies enacted in New York (NY) over time. Full article ">Figure 2
Granger causality test between station counts and policy counts for New York (NY). The p-value is 0.0065, which is below 0.05, so we reject the null hypothesis. There is significant relationship between charging infrastructures and EV incentivizing policies. Full article ">Figure 3
VAR results for New York (NY). Forecast percent change in station counts over time vs. actual percent change in station counts over time. Full article ">Figure 4
VAR results for other states. Forecast vs. actual percent change in station counts over time for (a) Washington, (b) Washington DC, (c) Texas, and (d) New Jersey. Full article ">Figure 5
p-values representing predictor ability for each of the policy types on new charging stations. Full article ">Figure 6
Forecast net change in registrations counts over time vs. true net change in registration counts over time for New York (NY). Full article ">

9 pages, 464 KiB

Open AccessProceeding Paper

Analyzing Patterns of Injury in Occupational Hand Trauma Focusing on Press Machines: A Registry-Based Study and Machine Learning Analysis

by Sarthak Pattnaik, Parita Danole, Sagar Mandiya, Ali Foroutan, Ghazal Mashhadiagha, Yousef Shafaei Khanghah, Khatereh Isazadehfar and Eugene Pinsky

Eng. Proc. 2024, 68(1), 61; https://doi.org/10.3390/engproc2024068061 - 29 Jul 2024

Viewed by 294

Abstract

Objectives: The aim of the project is to analyze the data of patients who have been admitted to the emergency room due to severe hand and palm injuries. Methods: we have used data visualization and statistical analysis to observe trends in various factors [...] Read more.

Objectives: The aim of the project is to analyze the data of patients who have been admitted to the emergency room due to severe hand and palm injuries. Methods: we have used data visualization and statistical analysis to observe trends in various factors pertaining to the patients, such as place of injury, machine-causing injuries, date and time of the injury, amputation, fracture, etiology, distribution of the injured hand, etc. Results: There is a significant difference between age and gender groups across various injuries. Most of the injuries in the dataset are occupational injuries caused by press machines. Most injuries take place in the later half of the week, on Wednesdays and Saturdays. Conclusion: There were 1676 patients who reported to the medical emergency center. Of these, only a handful of them have undergone extremely painstaking injuries where there was uncontrolled bleeding and hemi-amputation. We can also surmise the same by looking at the data that provides the summary of the number of fingers injured. Most patients have either one or two fingers injured. Very few patients had more than two fingers injured. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

9 pages, 936 KiB

Open AccessProceeding Paper

Decomposing the Sri Lanka Yield Curve Using Principal Component Analysis to Examine the Term Structure of the Interest Rate

by K P N Sanjeewa Dayarathne and Uthayasanker Thayasivam

Eng. Proc. 2024, 68(1), 62; https://doi.org/10.3390/engproc2024068062 - 27 Aug 2024

Viewed by 164

Abstract

In this study, we delve into the dynamics of the Sri Lankan government bond market, building upon prior research that focused on the application of principal component analysis (PCA) in modelling sovereign yield curves. Our analysis encompasses data spanning from January 2010 to [...] Read more.

In this study, we delve into the dynamics of the Sri Lankan government bond market, building upon prior research that focused on the application of principal component analysis (PCA) in modelling sovereign yield curves. Our analysis encompasses data spanning from January 2010 to August 2022. The study applied several PCA variants such as multivariate PCA, Randomized PCA, Incremental PCA, Sparse PCA, Functional PCA, and Kernel PCA on smoothed data. Kernel PCA was found to explain the majority of the variation associated with the data. Findings reveal that the first principal component accounted for a substantial 97.69% of the variations in yield curve movements, 2nd PCA accounted for 1.88%, and 3rd for 0.42%. These results align with previous research, which generally posits that the initial three principal components tend to elucidate around 95% of the fluctuations within the term structure of yields. Our results question the empirical findings, which state that the 1st PCA represents the longer tenor of the yield curve. In Sri Lanka, instead, the 1st PCA represents the 3-year bond yields. It may be because of the liquidity constraints in underdeveloped frontier markets, where longer tenor yields do not react fast enough to reflect the movement of the yield curve. The 2nd PCA represents the slope of the yield curve which is the yield difference of a 10-year T-Bond and 3 months T-Bill. The 3rd PCA which represents the curvature of the yield curve attributed to 2 × 3 years T-Bond yield—3 months T-bill10-year T-Bond. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

17 pages, 2569 KiB

Open AccessProceeding Paper

Oil Price Volatility and MENA Stock Markets: A Comparative Analysis of Oil Exporters and Importers

by Khalil Mhadhbi and Ines Guelbi

Eng. Proc. 2024, 68(1), 63; https://doi.org/10.3390/engproc2024068063 - 2 Sep 2024

Viewed by 511

Abstract

This paper explores the transmission of volatility from Brent oil price evolution to the stock returns of 7 MENA countries, encompassing three importers and four exporters, after excluding four initial countries using the ARCH test. Employing the GARCH-BEKK estimation method, we detect this [...] Read more.

This paper explores the transmission of volatility from Brent oil price evolution to the stock returns of 7 MENA countries, encompassing three importers and four exporters, after excluding four initial countries using the ARCH test. Employing the GARCH-BEKK estimation method, we detect this transmission from January 2008 to September 2022. The results reveal significant volatility persistence across six stock markets with three importer countries and three exporters. These findings align with Shiller’s theory, indicating high volatility in financial markets. Tunisia’s stock market shows sensitivity to oil market developments, while the Omani market demonstrates volatility transfer from Brent oil prices. However, Morocco’s market exhibits resilience, with no significant transmission from international oil prices. Exporting countries, except the UAE, display significant and positive coefficients, indicating volatility transmission. The study suggests further research into underlying mechanisms and recommends policymakers and investors implement strategies to mitigate volatility effects. Advanced modeling and behavioral insights can enhance risk management strategies. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

11 pages, 7567 KiB

Open AccessProceeding Paper

Towards an Automatic Tool for Resilient Waterway Transport: The Case of the Italian Po River

by Maria Luisa Villani, Ebrahim Ehsanfar, Sohith Dhavaleswarapu, Alberto Agnetti, Luca Crose, Giancarlo Focherini and Sonia Giovinazzi

Eng. Proc. 2024, 68(1), 64; https://doi.org/10.3390/engproc2024068064 - 4 Sep 2024

Viewed by 130

Abstract

Improved navigability can enhance inland waterway transportation efficiency, contributing to synchro-modal logistics and promoting sustainable development in regions that can benefit from the presence of considerable waterways. Modern technological solutions, such as digital twins in corridor management systems, must integrate functions of navigability [...] Read more.

Improved navigability can enhance inland waterway transportation efficiency, contributing to synchro-modal logistics and promoting sustainable development in regions that can benefit from the presence of considerable waterways. Modern technological solutions, such as digital twins in corridor management systems, must integrate functions of navigability forecasts that provide timely and reliable information for safe trip planning. This information needs to account for the type of vessel and for the environmental and geomorphological characteristics of each navigation trait. This paper presents a case study, within the EU project CRISTAL, focusing on the Italian Po River, of which the navigability forecast requirements of a digital twin are illustrated. Preliminary results to deliver navigability risk information were obtained. In particular, the statistical correlation of water discharge and water depth, computed from historical data, suggested that efficient forecast models for navigability risk, given some water discharge forecasts, could be built. To this aim, the LSTM (long-short-term-memory) technique was used on the same data to provide models linking water discharge and water depth predictions. Future work involves further testing these models with updated real data and integrating outcomes with climatic and infrastructure management information to enhance the accuracy of the risk information. Full article

(This article belongs to the Proceedings of The 10th International Conference on Time Series and Forecasting)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Eng. Proc., 2024, ITISE 2024

The 10th International Conference on Time Series and Forecasting

Other

Further Information

Guidelines

MDPI Initiatives

Follow MDPI