ARCHIVES

Weather prediction is an essential task in climate science and societal planning, impacting industries such as agriculture, transportation, and disaster management. Traditional forecasting methods often struggle to deliver accurate predictions due to the complexity and dynamic nature of weather systems. This paper proposes a machine learning approach to forecast temperature and humidity using historical weather data. Linear Regression and a deep learning model implemented via PyTorch were trained to recognize complex patterns in multivariate data. The models were evaluated for accuracy, mean squared error, and generalization. A Flask-based deployment demonstrates the model's capability for real-time forecasting. The results indicate that machine learning can significantly enhance the accuracy of short-term weather prediction models when trained on sufficient and relevant historical data.