Comparative Study of Machine Learning Models for Weather Prediction Using Historical Temperature Data

Authors: Meghan Mane, Umesh Shinde, Sheetal Popat, Pranjal Sancheti

Country: India

Full-text Research PDF File:   View   |   Download

Abstract: Accurate weather prediction is essential for informed decision-making in a variety of sectors, such as agriculture, disaster management, energy management, and others. The precision of weather forecasting is particularly optimized with the advancement of machine learning techniques. This paper presents a comparative analysis of weather prediction models using historical temperature data from Colaba, Mumbai, from 2017 to 2022. We compare the Root Mean Squared Error (RMSE) values of three algorithms: CatBoost, Stacked GRU, and BiLSTM GRU. CatBoost attains an RMSE of 1.41, closely followed by Stacked GRU at 1.47 and BiLSTM GRU at 1.54. Attention Networks are technically more useful compared to LSTM networks. Their combination with various Deep Learning algorithms is giving better results. The results not only highlight the relative performance of the model but also illustrate the vital role machine learning plays in enhancing weather prediction methods.

Keywords: Weather prediction, Attention model, Temporal Machine Learning model, Deep Learning.

Paper Id: 231403

Published On: 2022-05-04

Published In: Volume 10, Issue 3, May-June 2022

Cite This: Comparative Study of Machine Learning Models for Weather Prediction Using Historical Temperature Data - Meghan Mane, Umesh Shinde, Sheetal Popat, Pranjal Sancheti - IJIRMPS Volume 10, Issue 3, May-June 2022.

Share this