ANN Based Weather Analysis and Prediction

Abstract

Weather has always been an unpredictable factor and this affects the day to day lives of ordinary people. In order to make life easier it’s always a better way to know the weather conditions in advance, to act or plan the day or next accordingly. Many Artificial Intelligence (AI) techniques available can be used to train the model to predict the results. The more advanced the model, the more accurate are the results. In this model we are using LSTM and Bi–LSTM to predict the weather based on obtained datasets. The datasets are obtained from around the world, which means Australia, Bangladesh, Austin Texas, Jenna Germany. We preprocess the data before performing LSTM and Bi – LSTM algorithms on it and find out the accuracy through Mean Absolute Error and Root Mean Squared Error and compare the algorithms results in predicting the temperature.