Prediction of Rainfall in India: A Hybrid Model Approach

Abstract

Complexity of rainfall forecasting has been considered as an utmost research relevance in recent times. Precipitation is extremely beneficial in maintaining atmospheric balance. Rainfall is one type of precipitation. Though excessive rainfall harms the earth in a variety of ways, it is regarded as extremely valuable because it is one of the necessities for human survival. As a result, wise utilisation of rainfall water should be planned to minimise the drought condition and flood occurring in the country. Thus,a prediction of rainfall also forms a major part in planning things. This paper proposes a new hybrid model Moving average-kNN for doing the prediction. Error measures Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) have been used for the validity of the model.