Predicting Rainfall from Historical Data Trends

Abstract

Weather forecasting is and has been an attempt by meteorologists to predict the weather conditions with previous data trends recorded from the past.The weather forecasting is very important in countries like India where majority of the economy is dependent on agriculture. Rainfall prediction is the main factor we are studying and implementing here. Few algorithms used for rainfall prediction are LSTM, Genetic Algorithms and GRU, these are computer aided rule-based algorithms. In this paper, we performed qualitative analysis using GRU (Gated Recurrent Unit) algorithm.The GRU is the newer generation of Recurrent Neural networks and is pretty similar to an LSTM. GRU is also known to give accurate results and it is also good for processing sequence data for predictions.

Keywords: Weather Prediction, Data Mining, Weather Condition, Decision Trees