Forecast of Agricultural Crop Price using Time Series and Kalman Filter Method

This study is an exploratory analysis of the possibility of improving the accuracy and precision of typical time series model in forecasting future prices of rice crop by combining the techniques of ARIMA and Kalman filter respectively. Using actual rice data collected over a period of five years, the performance of the typical ARIMA model was compared to the combined performance of ARIMA-Kalman filter using Mean Square Error(MSE) and Root Mean Square Error(RMSE) as the bases of comparison. Results of the analysis revealed that a more accurate and precise time series estimates of future price of rice can be achieved when the technique of Kalman filter is combined with typical ARIMA time series model. Further analysis showed that predicted values generated by the combined techniques from out- of -sample forecasts are fairly closer to the actual values. On the basis of the findings of this study, the development of a time series software that will work on combining the algorithms of the ARIMA and Kalman filter is recommended.

Keyword: ARIMA, Kalman Filter, Forecast, Time Series .