Stochastic Modelling and Seasonal Artificial Neural Networks for the Prediction of Monthly Rainfall in India

Abstract

Stochastic modelling for the prediction of monthly rainfall levels with machine learning is dealt within this paper.Rainfall is essential for agriculture and the survival of all living things. Accurate forecasting of rainfall is essential for proper planning and the execution of agricultural activities. Periods of dry weather without rainfall can have major consequences on water supply affecting ground water levels and agriculture. Stochastic modelling has been widely used to time series prediction. This work presents a stochastic model for the monthly rainfall level prediction in India. Stochastic modelling based on Markov chains are used in this work. Second order Markov chain yields better results than first order. Seasonal artificial neural networks (SANN) give accurateprediction of the future rainfall level and determination of the prediction of monthly rainfall levels. The rainfall prediction is done on rainfall data in India provided by Indian Meteorological department. The predicted results obtained show the surpassing performance of the second order markov chain and seasonal artificial neural networks.

Keywords: Stochastic modelling, seasonal artificial neural networks, second order Markov chain, rainfall level, forecasting

Cite this Article

P. Arumugam, S.M. Karthik. Stochastic Modelling and Seasonal Artificial Neural Networks for the Prediction of Monthly Rainfall in India. Research & Reviews: Journal of Statistics. 2018; 7(1): 74–81p