组合预测模型在猪肉价格预测中的应用研究

Research on the Combinational Model for Predicting the Pork Price

本文在分析了神经网络、灰系统和时间序列预测模型的基础上,设计了将其中两种模型组合的预测方法。该方法的主要思想是利用回归预测思想将预测分为因素预测和结果预测两部分,并分别采用不同预测模型进行预测,从而达到提高预测精度的目的。利用该方法对吉林省近期的生猪价格进行预测,实验结果表明,该方法比单个预测方法有更好的预测效果,并且通过对不同组合的实验结果的分析发现,灰系统与神经网络相结合的方法具有更高的预测精度。

Based on grey systems, artificial neural networks, and time series, this paper devises a method of forecast combining two of the models which divide the forecast into two elements including factors and results by using regression. What is more, to improve the precision of the result, the method adopts different models for the factors and results predic- tion. The superior effect to each single method has been proved by the experimental result of predicting the recent price of pork in the province of Jilin. Furthermore, by analyzing different combinations, the method based on artificial neural net- works and grey systems is of better accuracy than others.