FORECASTING AGRICULTURAL PRODUCTS PRICES USING TIME SERIES METHODS FOR CROP PLANNIN

The cost of rice production in Thailand has continuously increased over time. To protect farmers from declining returns, there is a need for improving productivity of rice and use of crop rotation. This study aimed to apply predictive data analysis for rice cultivation planning to determine the suitable rotation crop from the following crops: turnips, muskmelons, kailan, peanuts, cantaloupes, and water mimosas under limited resources condition. Sixty-month selling data of Hom Pathum Rice and other six alternative crops was analyzed using the following time-series analysis methods: Least Square Method, Moving Average Method (3 months, 5 months, 7 months), Single Exponential Method, Double Exponential Method, and Winters' Method. Prediction efficiency was measured by the Mean Absolute Percentage Error (MAPE), Mean Absolute Deviation (MAD) and Mean Square Deviation (MSD). Then the suitable prediction method was used to forecast selling price of rice and rotation crops for 12 case studies under 3 scenarios. We found that case study 6 gave the highest profit (% increase when compared with a traditional method).