The Forecasting of Durian Production Quantity for Consumption in Domestic and International Markets

Abstract

Thailand is one of the world’s first exporting countries of fresh and processed durians. Each year in the durian season, there are excess supplies of fresh durians which directly cause a decrease in durian price. The farmers sell their durians at a price which is actually lower than the production cost. This problem recurs every year. The purpose of this research is to design and develop the models which can effectively forecast the quantity of fresh durian production. Firstly, applying the four Time Series models, secondly, applying the Back-propagation Neural Networks (BPN) model to find an accurate forecasting model that can effectively forecast the quantity of durians in Thailand in advance. The findings of the research are the model of Back-propagation Neural Networks of the structure 4-8-1 equal to the least value of Mean Absolute Percentage Error (MAPE). It is in the good level of forecasting, and can be applied to forecast the fresh durian quantity effectively. After attaining the accurate forecasting model, this is applied with the Linear Programming (LP) model to assess the value of appropriate fresh durian in each region in the following year. The data of fresh and processed durians planning can be helpful to the farmer, as they can then make the maximum profit from selling their durians