Weather Prediction Using Machine

Learning

Some information about Weather

Prediction

Sujata Saha
 Weather Prediction Using Machine Learning

Dr. B. C. Roy Engineering College, Durgapur

Submitted in partial fulfilment of the requirements for the award of

Bachelor OF Technology Degree in

Information Technology

By

SUATA SAHA

2503133

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

B. C. ENGINEERING COLLEGE

DURGAPUR – 713206, WEST BENGAL, INDIA

2
 OCTOBER 2025

B. C. ROY ENGINEERING COLLEGE

(Establish 21 August 2000 affiliated by MAKUT)
Jemua Road, Fuljhore, Durgapur – 713206, West Bengal
info@[Link]

COLLEGE OF COMPUTING

BONAFIDE CERTIFICATE

This is certify that this Professional Training Report is the bonafide work of
SUJATA SAHA (Reg. No. 2503133) who under the professional training in “WEATHER
PREDICTION under our supervision from AUGUST 2025 to DECEMBER 2025.

Internal Guide
Sabbir Reza Tarafdar ME

Head of the Department

Dr. Arindam Ghosh

Submitted for Viva voce Examination held

on____________________________________

3
Internal Examiner External Examiner

DECLARATION

I, SUJATA SAHA (Reg. No. 2503133), hereby declare that the professional Training Report on
“WEATHER PREDICTION USING MACHINE LEARNING” has been prepared by me under the
guidance of ____________________ at B. C. Roy Engineering College, Durgapur, and is
submitted in partial fulfilment of the requirement for the award of the Bachelor of Technology
Degree in Computer Science and Engineering.

DATE: 09/10/2025

PLACE:DURGAPUR SIGNATURE OF CANDIDATE__________________________

4
 ACKNOWLEGMENT

I am pleased to acknowledge my sincere thanks to the Board of Management of B. C. ROY

ENGINEERING COLLEGE for their kind encouragement in doing the project and for the
completing it successfully. I am grateful to them.

I convey my thanks to Prof. (Dr.) Narendra Nath Pathak, who also serves as the Principal, Dr.
Arindam Ghosh head of the department of cse and Technology for providing me necessary
support and details at the right time during the progressive reviews.

I would like to express my sincere and deep snse of gratitude to my project Sabbir Reza
Tarafdar , for his valuable guidance, suggestion and constant encouragement paved way for the
successful completion of my project work.

I wish to express my all thanks to all the Teaching and Non- Teaching Staff members of the
Department of Information Technology who were helpful in, many ways for the completion of
project.

5
 ABSTRACT
Accurate weather forecasting is critical for agriculture, disaster preparedness, transportation,
and daily life. Traditional meteorological models, while effective, often struggle with nonlinear
patterns and require extensive computational resources. Machine learning (ML) offers a
promising alternative by leveraging data-driven approaches to model complex atmospheric
dynamics. This study explores the application of ML algorithms—such as decision trees, support
vector machines, and deep neural networks—for predicting temperature, precipitation, and
extreme weather events. By training models on historical weather data and satellite imagery,
ML systems can identify hidden patterns and improve forecast accuracy. The integration of real-
time data streams and adaptive learning further enhances responsiveness to changing
conditions .This abstract highlights the transformative potential of machine learning in modern
meteorology, emphasizing its scalability, efficiency, and ability to complement traditional
forecasting techniques.

6
 TABLE OF CONTENTS

Chapter Page
No. TITLE No.

ABSTRACT
LIST OF FIGURES
LIST OF ABBREVISTIONS AND MODULES

1. INTRODUCTION

2. AIM AND SCOPE

3. PROJECT DESCRIPTION

7
4. SYSTEM ANALYSIS AND DESIGN

5. CONCLUSION AND FURURE WORK

REFERENCES
6. APPERENDIX

8
 CHAPTAR 1

INTRODUCTION

1.1 GENERAL

Weather forecasting plays a vital role in modern society, influencing sectors

such as agriculture, transportation, disaster management, and daily human
activities. Traditional forecasting methods rely heavily on complex
mathematical models and physical simulations of atmospheric behavior.
While these models have advanced significantly, they often struggle with the
nonlinear and chaotic nature of weather systems, especially when dealing
with large datasets and real-time predictions.
Machine learning (ML), a subset of artificial intelligence, offers a powerful
alternative by enabling systems to learn patterns from historical weather data
and make accurate predictions without explicitly programming the underlying
physics. ML algorithms such as regression models, decision trees, support
vector machines, and deep neural networks have shown promising results in
forecasting temperature, humidity, rainfall, and even extreme weather
events. These models can process vast amounts of data from satellites,
sensors, and weather stations, uncovering hidden correlations and improving
forecast precision.

1.2 OBJECTIVE
To develop and evaluate machine learning models for accurate and efficient
weather prediction by analyzing historical and real-time meteorological data,
aiming to enhance forecasting precision and support climate-related decision-
making.

9
1.3 OVERVIEW OF THE PROJECT
This project focuses on leveraging machine learning techniques to improve
the accuracy and efficiency of weather forecasting. Traditional meteorological
models often rely on complex physical equations and require significant
computational resources. In contrast, machine learning offers a data-driven
approach that can uncover hidden patterns in large volumes of historical and
real-time weather data.
The project involves collecting datasets from reliable sources such as weather
stations, satellites, and climate databases. These datasets include variables
like temperature, humidity, wind speed, and precipitation. Various machine
learning algorithms—such as linear regression, decision trees, random forests,
and neural networks—are trained and tested to predict future weather
conditions.

2. Methodology
2.1 Data Collection

Historical weather data is gathered from reliable sources such as meteorological departments,
satellite feeds, and open-access climate databases. The dataset includes variables like
temperature, humidity, wind speed, pressure, and precipitation, spanning multiple years and
regions to ensure diversity and accuracy.

2.2 Data Preprocessing

Raw data often contains missing values, noise, and inconsistencies. Preprocessing steps include:

 Handling missing data using interpolation or imputation techniques.

 Normalizing or scaling features for uniformity.
 Encoding categorical variables (e.g., weather conditions).
 Splitting data into training, validation, and test sets.

2.3 Model Selection

Various machine learning algorithms are selected based on the nature of the data and prediction
goals. These may include:

 Linear Regression for temperature prediction.

10
  Decision Trees and Random Forests for classification tasks.
 Support Vector Machines (SVM) for pattern recognition.
 Neural Networks for complex, nonlinear forecasting.

2.4 Model Training and Evaluation

Models are trained using the processed dataset and evaluated using performance metrics such as:

 Mean Absolute Error (MAE)

 Root Mean Square Error (RMSE)
 R² Score Cross-validation techniques are used to ensure model robustness and prevent
overfitting.

2.5 Deployment and Real-Time Forecasting

The best-performing model is deployed in a real-time environment using APIs or cloud

platforms. Live weather data is fed into the system to generate dynamic forecasts. The model is
periodically retrained with new data to maintain accuracy and adapt to changing climate patterns.

3. Experiment

The experiment involved applying various machine learning algorithms to historical weather
data to predict future conditions such as temperature, humidity, and rainfall. Data was collected
from trusted meteorological sources and preprocessed to handle missing values, normalize
features, and prepare it for model training.

Multiple models—including Linear Regression, Random Forest, Support Vector Machine

(SVM), and Long Short-Term Memory (LSTM) neural networks—were implemented using
Python libraries like Scikit-learn and Tensor Flow. The dataset was split into training, validation,
and testing sets to ensure fair evaluation.

Each model's performance was assessed using metrics such as Mean Absolute Error (MAE),
Root Mean Square Error (RMSE), and R² Score. Among the tested models, LSTM showed
strong performance in time-series forecasting, while Random Forest excelled in classification
tasks.

The experiment demonstrated that machine learning can significantly enhance the accuracy and
efficiency of weather prediction, especially when models are trained on diverse and high-quality
datasets.

4. Result

11
The machine learning models showed promising results in predicting weather parameters. LSTM
networks performed best for time-series forecasting, while Random Forest excelled in classifying
weather conditions. Evaluation metrics like MAE, RMSE, and R² confirmed high accuracy,
especially with well-preprocessed data. However, prediction of extreme events remained
challenging due to data imbalance. Overall, the experiment demonstrated that machine learning
can significantly enhance weather forecasting when trained on diverse and quality datasets.

5. Conclusion

This project demonstrates the potential of machine learning in enhancing the accuracy and
efficiency of weather forecasting. By analyzing historical and real-time meteorological data,
various ML models—especially LSTM and Random Forest—proved effective in predicting
temperature, rainfall, and other weather parameters. The results highlight that data-driven
approaches can complement traditional forecasting methods, offering faster and more adaptive
predictions. As climate patterns grow increasingly complex, integrating machine learning into
meteorological systems will be essential for informed decision-making across agriculture,
disaster management, and public safety.

6. References

1. Bochenek, B., & Ustrnul, Z. (2022).

Machine Learning in Weather Prediction and Climate Analyses—Applications and
Perspectives.
Atmosphere, 13(2), 180.
Available here
2. Mukherjee, A., Panda, J., Choudhury, A., & Giri, R. (2024).
Machine Learning and Deep Learning Applications in Weather and Climate Studies: A
Systematic Review.
Vayumandal Journal, Indian Meteorological Society.
Available here
3. Kumar, S., Raj, P., & Kumar, C. (2025).
Weather Forecasting Using Artificial Intelligence and Machine Learning.
International Journal of Science and Advanced Technology (IJSAT).
Available here

12