ISSN: 2959-6386 (Online), Vol.

2, Issue 1

Journal of Knowledge Learning and Science Technology

journal homepage: https://jklst.org/index.php/home

Leveraging Advanced Analytics for Reference Data Analysis in Finance

Manish Tomar1, Vathsala Periyasamy2
1
Citibank, USA.
2
Hexaware Technologies, USA.

Abstract

This paper explores the utilization of advanced analytics techniques for reference data analysis within the finance
sector. Reference data plays a crucial role in financial analysis, providing essential information for various processes
such as risk management, trading, and regulatory compliance. Leveraging advanced analytics methodologies enables
financial institutions to extract valuable insights from reference data, thereby enhancing decision-making processes
and improving operational efficiency. The paper discusses the challenges associated with reference data analysis in
finance and highlights the opportunities presented by advanced analytics approaches. By examining case studies and
industry best practices, the paper offers insights into how financial institutions can effectively leverage advanced
analytics to derive actionable intelligence from reference data.

Keywords: Advanced Analytics, Reference Data, Finance, Financial Analysis.

Article Information:
Article history: 13/06/2023 Accepted: 15/06/2023 Online: 30/06/2023 Published: 30/06/2023
DOI: https://doi.org/10.60087/jklst.vol2.n1.P136

Correspondence author: Manish Tomar

Introduction

Introduction:

In the realm of finance, reference data serves as the backbone of critical decision-making processes, underpinning risk
management strategies, trading activities, and regulatory compliance efforts. Reference data encompasses a wide array
of information, including securities identifiers, counterparty details, pricing data, and market reference points, among
Journal of Knowledge Learning and Science Technology ISSN: 2959-6386 (Online), Vol. 2, Issue 1129

others. Effectively harnessing and analyzing this wealth of reference data is paramount for financial institutions
seeking to gain a competitive edge in today's dynamic market landscape.

Traditionally, financial institutions have relied on manual processes and legacy systems to manage and analyze
reference data. However, with the advent of advanced analytics techniques, there exists a significant opportunity to
unlock deeper insights and extract greater value from this data. Advanced analytics methodologies, including machine
learning, natural language processing, and predictive analytics, offer sophisticated tools to uncover hidden patterns,
identify trends, and make more informed decisions based on reference data.

This paper delves into the role of advanced analytics in reference data analysis within the finance sector. It explores
the challenges faced by financial institutions in managing and analyzing reference data and examines how advanced
analytics approaches can address these challenges. By leveraging advanced analytics, financial institutions can
enhance their ability to extract actionable intelligence from reference data, thereby improving decision-making
processes, mitigating risks, and driving operational efficiency.

Throughout this paper, we will delve into case studies and industry best practices to illustrate how financial institutions
are leveraging advanced analytics to optimize reference data analysis. By understanding the benefits and opportunities
presented by advanced analytics in finance, organizations can position themselves for success in an increasingly
competitive and data-driven environment.

Literature Review:

Advanced analytics is being leveraged for reference data analysis in finance [1]. The use of technology and advanced
analytics in audits, specifically in the banking sector, has led to significant changes in the loan review process [2].
Auditors can now analyze the entire population of transactions in detail, saving time and resources [3]. However, it is
important to note that while technology can replace lower-level accounting and auditing skills, human ability to
understand and interpret business situations cannot be replaced [4]. The application of advanced analytics in finance
is not limited to audits. Big data analytics is becoming increasingly important in the banking, finance, and insurance
sectors [5]. These advancements have implications for audit committees and the adoption of useful big data approaches

Methodology

1. Data Collection: The first step in leveraging advanced analytics for reference data analysis involves collecting
relevant datasets from various sources. This may include internal databases, external data providers, market feeds, and
regulatory filings. The collected data should cover a wide range of reference data elements, such as securities
identifiers, pricing data, and counterparty information.

2. Data Preprocessing*: Once the data is collected, it undergoes preprocessing to ensure its quality and consistency.
This involves tasks such as data cleaning, normalization, and deduplication to remove errors, inconsistencies, and
redundancies from the dataset. Additionally, missing values may be imputed using techniques such as mean
substitution or predictive modeling.

3. Feature Engineering: Feature engineering is a crucial step in preparing the data for analysis. This involves selecting
and transforming relevant features from the raw data to create meaningful variables for analysis. Techniques such as
dimensionality reduction, aggregation, and transformation may be applied to extract valuable insights from the
reference data.
 1130 Manish
Tomar

4. Model Selection: With the preprocessed data and engineered features in hand, the next step is to select appropriate
analytics models for analysis. Depending on the specific objectives of the analysis, various machine learning
algorithms, such as regression, classification, clustering, and anomaly detection, may be employed to uncover patterns
and relationships within the reference data.

5. Model Training and Evaluation: Once the models are selected, they are trained on a portion of the data and evaluated
using metrics such as accuracy, precision, recall, and F1-score. Cross-validation techniques may be used to assess the
generalization performance of the models and mitigate overfitting.

6. Model Deployment: After the models are trained and evaluated, they are deployed into production environments
for real-time analysis of reference data. This may involve integrating the models with existing systems and workflows
to automate decision-making processes and provide actionable insights to stakeholders.

7. Monitoring and Maintenance: Once deployed, the models are continuously monitored to ensure their performance
remains optimal over time. This may involve tracking key performance indicators, detecting drifts in data distribution,
and retraining the models periodically to adapt to changing market conditions and regulatory requirements.

By following this methodology, financial institutions can effectively leverage advanced analytics to analyze reference
data, extract valuable insights, and make informed decisions to drive business success.

Background:

The theoretical foundation underlying decision support systems (DSS) and their integration within the broader
framework of Adaptive Case Management (ACM) underscores the dynamic nature of decision-making processes and
contexts. According to Fischer and Giaccardi (2004), key features of DSS include the development of socio-technical
environments to support users throughout system development and usage phases, fostering social creativity through
collaborative exchanges of ideas, integrating art and design for self-realization processes, and employing meta-
analysis techniques for synthesizing and generalizing previous studies.

ACM, as articulated by Fischer and Giaccardi, encompasses a collaborative approach to assessing, planning,
facilitating, and advocating for options and services tailored to meet individuals' holistic needs. This collaborative
process, facilitated by communication and available resources, aims to achieve quality cost-effective outcomes. Under
this definition, ACM serves as a platform that incorporates a decision support system, as depicted in Figure 1.

Clyde Holsapple, a pioneer in DSS, emphasized that DSS architecture serves as an ontology, providing a common
framework rather than defining the essence of DSS itself.
Journal of Knowledge Learning and Science Technology ISSN: 2959-6386 (Online), Vol. 2, Issue 1131

The language for the design, discussion, and evaluation of Decision Support Systems (DSS) as articulated by
Holsapple (2008) and Adam (1996) provides valuable insights into the architecture and functionality of DSS.
Holsapple defines DSS architecture as a framework that identifies the essential elements of a DSS and their
interrelationships, while Adam emphasizes the relationships between system components and business goals in the
context of Adaptive Case Management (ACM) systems.

Research conducted by various authors highlights the common objectives of DSS, including supporting knowledge
workers in making optimal decisions, facilitating faster and more accurate case resolution, and enhancing business
agility by adhering to established business rules. Control functions within DSS are typically executed through meta-
 1132 Manish
Tomar

knowledge subsystems, which incorporate norms, axioms, and ontologies to facilitate adaptive and generative learning
processes.

ACM system users are instrumental in building corporate knowledge through the utilization of IT tools and social
mechanisms, enabling the dissemination of tacit knowledge across the organization for more effective case processing.
It is crucial for DSS projects to analyze interactions among business stakeholders, viewing managers as active agents
within the Decision Making Network (DMN) as described by Perry (2014).

ACM can be conceptualized as an IT platform integrating a decision support system, with the DMN serving as a
process map that visualizes all possible case states and provides process managers with comprehensive insights into
business operations. Individuals dealing with less structured decision problems must possess a solid understanding of
the problem-solving process and relevant techniques to utilize system resources effectively, ensuring optimal decision
outcomes.

Extending the Decision Support System (DSS)

Extending the Decision Support System (DSS) model to incorporate big data analysis into Adaptive Case Management
(ACM) presents numerous opportunities for organizations to enhance decision-making processes and improve
operational efficiency. ACM systems, which automate document-intensive business processes, can both serve as a
data source for big data applications and benefit from big data analytics to optimize decision-making.

Big data analytics can transform case management applications by automating human decisioning processes and
optimizing them for efficiency. Human decisioning processes are often prone to errors and inconsistencies, especially
during periods of increased workload, and are heavily reliant on staff training. Leveraging big data analytics can help
uncover insights that individual case workers might overlook and identify trends based on historical and current data,
ultimately making the decision-making process more intelligent and informed.

The synergy between big data and ACM holds significant potential for organizations to derive actionable insights
from vast amounts of data. Big data, characterized by its high volume, velocity, and variety, presents both challenges
and opportunities for organizations to extract value and drive decision-making. By harnessing predictive analytics and
other advanced methods, organizations can make more confident decisions that lead to greater operational efficiency,
cost reduction, and risk mitigation.

The collaboration between ACM and big data is particularly beneficial for small to medium-sized enterprises, where
ACM tools are chosen for their flexibility and shorter implementation cycles. However, in larger companies with
multiple IT departments and business unit silos, the lack of synergy and integration among applications may result in
inefficiencies and increased IT costs. Nonetheless, the big data trend opens up new data sources and opportunities for
organizations to analyze processes deeply and simulate potential improvements.

Successful implementation of big data analytics in ACM requires careful planning and collaboration among decision-
makers in IT and business, as well as engagement with subject domain experts. ACM enables business knowledge
workers and processes to adapt to changing situations, continuously refine business performance drivers, and make
better decisions. The natural affinity between big data and cloud computing further facilitates the accessibility of these
benefits to organizations of all sizes.
Journal of Knowledge Learning and Science Technology ISSN: 2959-6386 (Online), Vol. 2, Issue 1133

Results

Organizations and Problem Areas

The study focused on organizations within the healthcare sector, specifically Municipal Hospital Katowice and
Municipal Hospital Sosnowiec. These organizations have implemented Adaptive Case Management (ACM) as part
of their business operations. The aim was to explore the challenges and benefits associated with implementing ACM
and how it reshapes the organizations' approaches to information management.

Scope of the Studies

The scope of the studies aimed to describe the development of the organizations, highlighting the business challenges
and benefits resulting from the implementation of ACM. Special emphasis was placed on the organizations'
approaches to information management.

Case 1 – Monitoring Operating Conditions of Refrigerated Storage Equipment in Hospitals

Problem Definition: Monitoring the operating conditions of refrigerated storage equipment is essential for ensuring
the safety of patients in the neonatal pathology department of a research hospital in Poland. This involves monitoring
ambient temperature conditions under which medicinal products or food are stored.

Research Methods and Data Sources

The study utilized the observation method and examined various documents, including internal documents, vendor
factsheets, and software specifications. The selection of documents and case studies was based on their subject-matter
value. Qualitative data analysis was conducted using measurement transducers designed by s4bi sp. z o.o., which
measure temperatures with digital temperature sensors. The transducers transmit data to an MMC system, where
historical measurements are stored.

Key ACM Area

The study focused on incidents related to temperature variations in refrigerated storage equipment. Big data analysis
was employed to define reference values for equipment operation based on temperature distribution profiles over 24-
hour periods. Reference values were determined using basic mathematical statistics and parameter estimation, with
deviations from these values documented as incidents. The documentation formed the basis for developing problem
prevention procedures, leading to ongoing improvements in departmental work quality.
 1134 Manish
Tomar

Case 2 – Predictive Analysis and Maintenance

Problem Definition

The predictive maintenance approach was applied to predict the lifetimes of X-ray tubes in computed tomography
(CT) scanners at one of Poland's largest research hospitals. This was prompted by the high prices of X-ray tubes and
legal requirements mandating hospitals to undergo public procurement processes for purchases exceeding a certain
amount. The decision support system, based on the Adaptive Case Management (ACM) model with big data analysis,
aimed to schedule CT tests, plan purchase budgets, and define start dates for public contract processes to minimize
idle times of CT scanners.

Data Sources

The data comprised DICOM test pictures generated daily before tests, typically CT pictures of manikins with
predefined parameters such as dimensions, shape, and material. Each picture contained a header file with details like
the date of capture, scanner parameters, settings, and picture-specific information. The picture data was stored on
DICOM servers, supplemented by information on scanner use and failure incidents.

Processing Models

Upon completing a manikin-based test, the system computed quality parameters such as signal-to-noise ratio, grey
level contrast, and edge blurring ratios using the test pictures. These ratios, along with picture header file data, were
transmitted to the MMC system. The system also stored information on failure incidents and inspections. Due to the
large volume of data, specific tools and methods for big data processing were employed for optimized analysis. A
statistical model for X-ray tube operating times was developed using the multivariate survival modeling approach. A
Cox single-variable model was formulated for each parameter, followed by a Wald test to assess parameter
significance. Based on statistically significant parameters, a Cox multivariable proportional hazards model was
developed. Monitoring quality parameters allowed users to identify signs of X-ray tube malfunction, such as
increasingly blurred scans, enabling the prediction of replacement times with acceptable accuracy.

Discussion

Case 1

The case study highlights how Adaptive Case Management (ACM) platforms not only optimize organizational IT
infrastructure but also reshape operational strategies. By integrating ACM with Big Data analysis, companies can
enhance core operations, particularly in terms of accuracy and securing critical operational areas. ACM systems
facilitate improved knowledge management and decision support, leading to better outcomes.

Integrating Big Data analysis into the DSS model provided by ACM offers the potential to enhance decision-making
processes based on factual data rather than intuition or internal knowledge. When ACM users have access to accurate,
up-to-date information and analytical tools, organizations benefit from more informed decisions and consistent
Journal of Knowledge Learning and Science Technology ISSN: 2959-6386 (Online), Vol. 2, Issue 1135

outcomes. This approach transforms how knowledge workers execute their roles and influences overall business
performance.

Case 2

Organizations employing ACM principles effectively blend innovation with core operations, mitigating risks
associated with business process optimization. ACM enables dynamic process changes without chaos, allowing for
continual enhancement and adaptation of business processes based on validated knowledge. The case study also
suggests that leveraging Big Data analysis within a case management infrastructure enables proactive decision-
making, particularly in scenarios such as predictive maintenance, resulting in better-informed decisions.

Conclusion
The integration of ACM with Big Data analysis represents a modern business model capable of meeting the diverse
needs of contemporary organizations. By enhancing decision support systems with Big Data analysis, organizations
can make data-driven decisions, leading to improved operational efficiency and business outcomes.

ACM serves as a dynamic management strategy, enabling organizations to respond to evolving customer expectations
and market demands. The synergy between ACM and Big Data analysis empowers workers and enhances productivity
by leveraging accumulated skills and insights. This combination fosters a learning organization culture, facilitating
continual improvement in internal processes and driving superior business outcomes.

Overall, ACM, coupled with Big Data analysis, offers a promising platform for modern case management, enabling
organizations to adapt to changing business environments and achieve sustainable growth. The results of early Big
Data analytic projects underscore the potential for significant business value creation, making it imperative for
organizations to embrace this approach to remain competitive in today's market landscape.

References List:

[1].Adam, F. (1996). Exploration of Organisation Analyser: A Tool for Studying Decision-Making

Networks inOrganizations. In P. Humphreys, L. Bannon, A. McCosh, P. Migliarese, & J.-C. Pomerol
(Eds.), Implementing Systems to Support Management Decisions (pp. 1–20). London: Chapman & Hall.

[2]. Drucker, Peter F. (1999). Management Challenges for the 21st Century. New York, NY: Harper
Collins. (p.157).

[3]. Fischer, G., & Giaccardi, E. (2004). Meta-Design: A Framework for the Future of End-User
Development. In H. Lieberman, F. Paterno, & V. Wulf (Eds.), End User Development – Empowering
People to Flexibly Employ Advanced Information and Communication Technology (p. 224). Dordrecht,
The Netherlands: Kluwer Academic Publishers.

[4]. Holsapple, C. W. (2008). DSS Architecture and Types. In F. Burstein & C. W. Holsapple (Eds.),
InternationalHandbooks on Information Systems. Handbook on Decision Support Systems 1 – Basic
Themes (pp. 163– 189). Berlin-Heidelberg: Springer.

[5]. Perry, A. (2014). The Value of Adaptive Case Management in the Experience Economy. Thrive:
Blog for
1136 Manish
Tomar

[6]. Finance Professionals, August 8th. Retrieved from https://origin-

www.fpablog.lexmark.com/us/2014/08/08/the-value-of-adaptive-case-management-in-the-experience-
economy/

[7]. Van der Alst, V. M. P., & Berens, P. J. S. (2001). Beyond Workflow Management: Product-Driven
Case Handling. In S. Ellis, T. Rodden, & I. Zigurs (Eds.), International ACM SIGGROUP Conference
on Supporting Group Work (GROUP 2001) (pp. 42–51). New York, NY: ACM Press.

[8]. White, M. (2009). Delivering Case Management with BPM in the Public Sector: Combining
Knowledge withProcess. In L. Fischer (Ed.), 2009 BPM and Workflow Handbook: Spotlight on
Government (pp. 65–80). New York, NY: Future Strategies Inc.
[9]. Vemuri, N. V. N. (2023). Enhancing Human-Robot Collaboration in Industry 4.0 with AI-
driven HRI. Power System Technology, 47(4), 341-358. Doi: https://doi.org/10.52783/pst.196
[10]. Vemuri, N., Thaneeru, N., & Tatikonda, V. M. (2023). Smart Farming Revolution:
Harnessing IoT for Enhanced Agricultural Yield and Sustainability. Journal of Knowledge
Learning and Science Technology ISSN: 2959-6386 (online), 2(2), 143-148.
DOI: https://doi.org/10.60087/jklst.vol2.n2.p148