Prediction of Green Chilli production in Jaffna, Sri Lanka: ARIMA Approach

Abstract

Green chilli is one of the most important cash crops grown in Sri Lanka. Also, it has become an essential ingredient in Sri Lankan meals. Observing the fluctuation of Green chilli production is essential in the market economy. The level of the production and its’ fluctuation not only has a significant influence on farmers’ income and consumer price, but a reasonable and stable production also has an irreplaceable effect on the safe running of the chilli market. The green chilli production is made under two seasons, Yala and Maha. In this research, the annual production of green chilli in Jaffna was modeled by using Autoregressive Integrated Moving Average (ARIMA) approach. The data were obtained from the Department of Census and Statistics, during the period of 2002 to 2018. The Box-Jenkin’s procedure was used to build the ARIMA model for the annual chilli production. Further, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and Mean Squared Error (MSE) were used to select the best model among some possible models. Through modeling, it was identified that ARIMA (1,1,0) is the best fitting model for the given data. Moreover, the adequacy of the best model was examined using L’jung-box test. Finally, the identified best model was used for short term prediction.