Applying Artificial Neural Network to Predict the Volume of Crops in the Auction Market

• Main Authors: Hung, Wen-Ming, 洪文明
• Other Authors: Chen Deng-Neng
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/84722636600497324172

碩士 === 國立屏東科技大學 === 資訊管理系所 === 104 === The agricultural issue is always the critical factor to achieving the sustainability of a country. Due to the growing of plants is affected by weather, season, and lots of external influences, the harvest of crops is unstable and might cause the imbalance between supply and demand in the market. Therefore, the precise prediction of demand in the crops transaction market is important. It is helpful to the farmers to make the cultivating plan and also beneficial to the development of agriculture. In our research, we apply back propagation neutral network (BPNN) to develop a time-series prediction model. The model is used to predict the trading volume of banana and cabbage in the fruit transaction market in Taiwan. The trading volume indicates the demand in the market, for that reason, the farmers and government can make cultivating plan effectively by the prediction results. We collected the price and weather information of banana and cabbage from 2004 to 2014. The price and weather information were used to train the BPNN model and evaluate the accuracy of prediction. The analysis results showed the price of models have lower than 20% error rate. It implicates that BPNN can be used to predict the trading volume of crops in the market.