ARIMA and State-Space models for sugarcane (Saccharum officinarum) yield forecasting in Northern agro-climatic zone of Haryana

Abstract

Advance estimates of significant cereal and commercial crops are given by the Directorate of Economics and Statistics and the central Ministry of Agriculture, Cooperation & Farmers’ Welfare. However, the final estimates are released a few months after the actual harvest of the crops. In this study, ARIMA and State-Space models have been developed for sugarcane yield forecasting in Ambala and Karnal districts of Haryana. The above-mentioned models have been developed using yield data of sugarcane crop for the time period 1966-67 to 2009-10 of Ambala and Karnal districts. The validity of fitted models has been tested over the years 2010-11 to 2016-17. The forecasting performance of the developed models has been studied using percent deviations of sugarcane yield forecasts in relation to the actual yield, and root means squared errors. It has been observed that state-space models outperform the popular ARIMA models for forecasting of sugarcane yield in Northern Agro-climatic Zone of Haryana.