Arima model for and forecaste on tea production in India

Abstract

The paper explains the autoregressive integrated moving average (ARIMA) (Box-Jenkins) model in detail, develops an ARIMA model for tea production in India using monthly data for the period Jan. 1979 through July 1991, and applies the so developed model to forecast tea production during the future 12 months. On verification through several statistical techniques including the accuracy of sample period forecasts, the model is found to be quite strong. The post sample period forecasts are consistent with the seasonal pattern of the data in past 12 years and this enhances the modelzs credibility. As per the results, tea production in India during the post-sample one year is expected to be fluctuating between a low of 14,300 tonnes during Feb.-March 1992 to 90,600 tonnes in July 1992; the actual data for these periods are not yet available in published form.