Implementation of Long Short Term Memory Algorithm for Rainfall Prediction of East Madhya Pradesh, India
Author : Nisha Thakur, Dr. Shanti Swarup Dubey and Ravi Shrivastava
Abstract :
Rainfall is an important factor in Madhya Pradesh state as the economy is dependent on agriculture here. Time Series forecasting approach of univariant data for monthly rainfall prediction is done using Long Short Term Memory [LSTM]. To judge the accuracy of models two parameters were chosen one for error and another for accuracy i.e. Root mean square error (RMSE) and Cosine similarity (CS). Long short term memory model applied on 1404 monthly data of east Madhya Pradesh state. Various epoch(s) such as 15, 30, 45, 60, 75 and 90 respectively are done for LSTM approach. The computed value of RMSE was found as 0.1142, 0.0895, 0.1042, 0.1042, 0.0612, 0.0798 and 0.0673 respectively. Likewise, value of CS was found as 0.9628, 0.9778, 0.9762, 0.9812, 0.9776 and 0.9795 respectively. The experimental results show that Long Short Term Memory gave significant results in lower epochs.
Keywords :
LSTM, Bi-LSTM, deep learning, prediction, rainfall
