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

Making educated decisions in a number of fields, including agriculture, disaster relief, energy management, and others, depends on accurate weather forecasting. With the development of machine learning algorithms, weather forecasting accuracy is maximised. Using historical temperature data from Colaba, Mumbai, from 2017 to 2022, this research compares various weather forecast models. Three algorithms— CatBoost, Stacked GRU, and BiLSTM GRU—have their Root Mean Squared Error (RMSE) values compared. With an RMSE of 1.41, CatBoost comes in first, followed by BiLSTM GRU at 1.54 and Stacked GRU at 1.47. In theory, attention networks are more beneficial than LSTM networks. Better outcomes are being obtained when they are combined with different Deep Learning algorithms. The outcomes demonstrate the model's relative performance as well as the critical role machine learning plays in improving weather prediction techniques.