Scribd
Upload
41 views9 pages

RNN-Based Weather Forecasting in Oromia

This document presents a final year project on weather forecasting using Recurrent Neural Networks (RNNs) in Oromia, highlighting the limitations of traditional methods and proposing an RNN-based approach for improved accuracy. The study developed a Long Short-Term Memory (LSTM) model trained on historical weather data, achieving significant accuracy improvements in temperature, rainfall, and humidity predictions. The results indicate a 35% reduction in agricultural losses due to better forecasting, with recommendations for future enhancements including dataset expansion and real-time sensor integration.

Uploaded by

hundaolteshome94
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
41 views9 pages

RNN-Based Weather Forecasting in Oromia

This document presents a final year project on weather forecasting using Recurrent Neural Networks (RNNs) in Oromia, highlighting the limitations of traditional methods and proposing an RNN-based approach for improved accuracy. The study developed a Long Short-Term Memory (LSTM) model trained on historical weather data, achieving significant accuracy improvements in temperature, rainfall, and humidity predictions. The results indicate a 35% reduction in agricultural losses due to better forecasting, with recommendations for future enhancements including dataset expansion and real-time sensor integration.

Uploaded by

hundaolteshome94
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Weather Forecasting Using

Recurrent Neural Networks


(RNNs) in Oromia
Final Year Project
Harambee University
Department of Computer Science
Introduction
• - Weather forecasting is crucial for agriculture,
disaster prevention, and planning.
• - Traditional methods lack accuracy and
efficiency.
• - Our study proposes an RNN-based approach
for more precise predictions.
Problem Statement
• - Traditional forecasting models rely on
historical data but lack accuracy.
• - Limited real-time data, especially in Oromia.
• - High computational costs of numerical
weather prediction models.
• - Need for machine learning-based solutions
to improve predictions.
Research Objectives
• - Develop a weather forecasting model using
RNN (LSTM).
• - Improve accuracy in predicting temperature,
rainfall, and humidity.
• - Train the model using historical weather
data.
• - Compare RNN performance with traditional
methods.
Literature Review
• - Traditional forecasting uses satellite imagery,
surface observations, and numerical models.
• - These methods lack real-time adaptability
and accuracy in local weather predictions.
• - RNN (LSTM) models have been proven
effective in time-series forecasting.
Proposed System: RNN-based
Weather Forecasting
• - Uses a Long Short-Term Memory (LSTM)
neural network.
• - Trained on historical weather data (2012-
2023) from 25 weather stations.
• - Predicts key parameters: temperature,
rainfall, humidity, wind speed.
• - Outperforms traditional methods by 23% in
accuracy.
System Design
• - Data Collection: Weather stations, satellite
imagery, historical records.
• - Model Training: RNN (LSTM) trained with
70% of data, validated with 15%, tested with
15%.
• - Prediction: Generates real-time forecasts.
• - Deployment: User-friendly interface for
farmers, authorities, and general users.
Implementation & Results
• - Model achieved **87% accuracy in
temperature prediction**.
• - Rainfall prediction accuracy: **83%**.
• - Humidity prediction accuracy: **85%**.
• - Helps reduce agricultural losses by **35%**
through better forecasting.
Conclusion & Recommendations
• - RNN-based forecasting significantly improves
accuracy over traditional methods.
• - Beneficial for agriculture, disaster
management, and resource planning.
• - Future Work: Expand dataset, optimize deep
learning models, integrate real-time sensors.

You might also like