Two-level multivariate fuzzy logic based integrated model for monsoon rainfall prediction

Abstract

Rainfall prediction is considered as one of the complex tasks in the area of time series forecasting. Timely and accurate forecasting is essential for better decision making. Autoregression based methods have been widely used for rainfall forecasting. But not much works attempted Vector autoregression in this area. This paper presents an integrated multivariate rainfall prediction scheme using Vector autoregression coupled with fuzzy inference and similarity based measure. Proposed method is a two step process where output of first step is fed into the process in the next step. Result shows improved accuracy of the proposed fuzzy logic based model over the individual statistical model.