PATTERN RECOGNITION AND PERFORMANCE COMPARISON OF REGRESSION MODELS FOR WEATHER PREDICTION

In this paper, a machine learning based weather forecast system is proposed by utilizing a real-world weather data set as a case study to demonstrate the performance of the models. The proposed system is implemented by using the Python numpy, pandas libraries, pickle module and a few others. The aim is to predict parameters from the dataset using machine learning and also compare the performance of three different models (LSTM, GRU and SimpleRNN) on a time stamped raw data collected using sensors, recorded on an hourly basis and obtained from ISRO.