Use of Statistical Methods for Studying Price Behaviour of Arecanut Application of ARIMA Forecasting Model Nagashree N., Gracy C. P., Nayak Akshata Department of Agricultural Marketing, Cooperation and Business Management, University of Agricultural Sciences, GKVK, Bangalore- 560 065, Karnataka Online published on 9 October, 2012. Abstract The high volatility in the prices of high-value agricultural commodities like arecanut over the past one decade has necessitated accurate price forecasting for taking selling decisions by the farmers. In this paper, a forecasting approach has been described and performance of Autoregressive Integrated Moving Average (ARIMA) models has been examined through prediction accuracy using monthly data from October 2002 to July 2012 from a leading arecanut market of the country, viz. Shimoga in Karnataka. Some, alternative models, viz Akaike Information Criterion (AIC), Mean Absolute Percentage Error (MAPE) and Schwarz's Bayesian Information Criterion (SBC) have also been evaluated based on their power of forecasting. On validation of the forecasts from these models, ARIMA (0, 1, 0) model has been found to perform better than the other models. It has been found that models with least Mean Absolute Percentage Error (MAPE) of 5, 5, 6 could be successfully applied to forecast the prices of three varieties of arecanut, viz. Saraku, Bette and Edi, respectively. The forecast accuracy of the three price series ranged between 94 and 99 per cent for the concerned months. The paper has suggested that this model can facilitate effective decision-making by the farmers and stakeholders in selling their produce. Top Keywords Arecanut, ARIMA, MAPE, price forecasting, Karnataka. Top |