Rainfall Forecasting Using Optimized Support Vector Regression Technique

Abstract

Abstract—Rainfall prediction has been an important area of research in the field of hydrology. In this article an attempt is made to apply Support vector regression (SVR) in annual rainfall forecasting and comparing its performance with other models like Random forest regression (RFR) and Artificial Neural Network (ANN) model. This study proposes an approach by combining grid search technique with SVR model. Grid search technique carefully tunes the SVR’s hyper parameters and then SVR model is built using the tuned optimal parameters. Past hundred years rainfall data for the state of Meghalaya, India, are been used for the experimentation purpose. The experimental result shows that a well-tuned SVR model outperforms RFR and ANN model in predicting overall annual rainfall. 

Keywords— Rainfall prediction, grid search, artificial Neural Network (ANN), Support Vector Regression (SVRR), Random Forest Regression (RFR).

Cite this Article

Amitabha Nath, Sushweta Kar, Goutam Saha. Rainfall Forecasting using Optimized Support Vector Regression Technique. Journal of Web Engineering & Technology. 2019; 6(2): 29–33p.