浙江农业学报

• 生物系统工程 • 上一篇    下一篇

基于广义回归神经网络的蔬菜市场日价格预测

  

  1. (1上海理工大学 光电信息与计算机工程学院,上海200093;2上海商学院 财经学院,上海200235)
  • 出版日期:2015-07-25 发布日期:2015-08-03

Forecasting of vegetable daily price based on general regression neural network#br#


  1. (1 School of Optical\|Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China; 2 Faculty of Finance and Accounting, Shanghai Business School, Shanghai 200235, China)
  • Online:2015-07-25 Published:2015-08-03

摘要: 蔬菜市场日价格预测一直是研究的难点。根据GRNN原理,基于2010年12月13日至2014年4月18日的青菜市场日价格(合计799组数据),建立GRNN模型,并依此预测2014年4月7日至2014年4月18日的青菜市场日价格。结果表明:平均相对误差的绝对值为312%,最大绝对误差为024元,误差很小,且建立的模型在青菜市场日价格预测中具有良好的泛化能力。GRNN模型是进行蔬菜市场日价格预测的合适的神经网络模型。

关键词: 市场价格, GRNN, 价格预测, 蔬菜

Abstract: Accurate forecasting of vegetable daily price is difficult. In the present study, the daily price of green vegetables in market from December 13, 2010 to April 18, 2014 (total 799 samples) were collected, and a GRNN model was established to predict market price of green vegetables from April 7, 2014 to April 18, 2014. The results showed that the absolute value of average relative error was 312%, and the maximal absolute error was 024 yuan, which was small. Thus, the estimated model for daily price forecasting of market vegetables exhibited good generalization ability, and was an appropriate neural network model.

Key words:  market price;GRNN;price forecasting, vegetable