Prediction of Temperature in Indian Metropolitan Cities Using Linear Regression and Long Short Term Memory Models

UHI is the process whereby human activity and urban infrastructure cause urban regions to suffer from warmer temperatures than rural ones. This study explores the use of machine and deep learning models for temperature prediction based on the Urban Heat Island (UHI) phenomenon. The study used historical data on UHI and temperature for a specific urban area. Preprocessing techniques were applied on the data and the model was trained and tested using MLR and LSTM. A comparative analysis of the models performance was done wherein it was observed that MLR outperforms LSTM in prediction of temperature. The MLR model is a classical linear regression model that assumes a linear relationship between temperature and relevant environmental variables such as humidity, pressure, wind speed, and direction. The LSTM model, is an algorithm that uses a recurrent neural network architecture to capture temporal dependencies and long-term patterns in the temperature data. The study also analyzes the feature importance of the environmental variables used in the MLR algorithm and the importance of the time-series data in the LSTM model. The results show that temperature and humidity are the prominent features in the MLR model, while the LSTM algorithm heavily relies on time-series data.