Rainfall Forecasting using Non-Parametric Bayesian Approach

Rainfall forecasting is very important to avoid many natural calamities like floods. Flooding is a dangerous natural calamity and it affects various geographic and economic sectors. It is one of the three major natural disasters in India, the other being drought and cyclone. As the world is getting intelligent, new techniques are being adopted for accurate prediction and prevention of these disasters. Flooding occurs mainly due to improper water management and sudden and unexpected rainfall. Rainfall is affected by many statistically varying parameters which are not finite. Here a non-parametric Bayesian technique, Gaussian process regression is adopted to predict the rainfall pattern.