A Research on Development of a Fuzzy Price Forecasting System for Vegetables and Fruits

研究了批发市场、农业合作社和农场价格预测的必要性，开发了蔬菜水果价格模糊预测系统的原型。结果如下：蔬菜和水果的批发市场在货物到达市场之前就开始定价，因为市场试图只把一些蔬菜卖给那些提前一天订购的人，或者出于其他一些原因。因此批发市场需要价格预测。农业合作社需要价格预测。容忍水平为<正负10%。鸟取县生产山药的农场也需要价格预测。误差范围从<正-负> 8%到<正-负> 20%，本研究开发的模糊价格预测系统原型有步骤。第一步，采用模糊线性回归方法对参数为模糊数的模糊线性供需模型进行估计，并对模型的一年后批发价格进行预测。隶属函数的形式对价格预测有影响。第二步，采用模糊推理的方法对第一步的预测价格进行修正。这一步需要一个月后的天气信息和其他生产中心的种植面积。如果收集得当，可以纠正第一步预测的价格。

The necessities of price forecasting of wholesale markets, agricultural co-operatives and farms were investigated and a prototype of fuzzy price forecasting system of vegetables and fruits was developed in this research. The results were as follows.Wholesale markets for vegetables and fruits have come to price their goods before they arrive at the markets, because the markets try to sell some vegetables only to those who have ordered a day in advance, and for some other reasons. Therefore wholesale, markets need price forecasting.Agricultural co-operatives need price forecasting. The tolerance levels are <plus-minus>10 percents. The farms who produce Chinese yams in Tottori Prefecture need price forecasting, too. The tolerance levels are from <plus-minus>8 percents to <plus-minus>20 perA prototype of fuzzy price forcasting system that was developed in this reseach has steps. In the first step, the fuzzy linear demand and supply models that have parameters of fuzzy numbers are estimated by fuzzy linear regression, and wholesale price after one year is forecasted means of these models. The form of membership function has an effect on price forecasting. In the second step, the forecasting price in the first step is corrected by means of fuzzy reasoning. The information of wather after one month and the planting area of the other production centers is needed in this step. If they are collected properly, the price that is forecasted in the first step can be corrected.