基于GM(1,1)模型和BP人工神经网络模型的蚕茧价格预测方法  

Forecasting of Cocoon Price Based on GM(1,1) Model and BP Artificial Neural Network Model

作　　者：周逸斌 黄凌霞[1] 

机构地区：[1]浙江大学动物科学学院,杭州310058

出　　处：《蚕业科学》2015年第6期1102-1107,共6页ACTA SERICOLOGICA SINICA

基　　金：现代农业产业技术体系建设专项(No.CARS-22);国家科技支撑计划项目(No.213BAD20B20);浙江省蚕桑创新团队项目(No.2011R50028)

摘　　要：应用灰色系统中的GM(1,1)模型预测蚕茧价格趋势,应用BP人工神经网络模型基于蚕茧产量波动预测蚕茧价格波动,建立组合2种模型预测蚕茧价格的方法,为蚕桑生产计划及蚕茧收售提供参考。应用该方法分别对2011—2014年各年度的蚕茧价格进行预测与检验,预测价格与实际价格的相对误差分别为4.17%、5.01%、2.84%和1.25%;应用该方法预测2015年的全国家蚕鲜茧销售均价为38.21元/kg,比上年度小幅增长。预测及检验结果表明,组合2种模型的蚕茧价格预测方法具有较高的准确度,适用于市场经济规律调控下的蚕茧价格预测。In present study, GM （1,1） prediction model in grey system was used to forecast cocoon price trend, and the BP artificial neural network model was used to forecast cocoon price fluctuation degree based on cocoon yield fluctuation. Furthermore, a method for forecasting of cocoon price was built combined with these two models to provide reference for sericulture production and cocoon sale and purchase. The annual cocoon prices of 2011 to 2014 were forecasted and verified by using this established method. The results showed that the relative errors between forecasted prices and actual prices were 4. 17%, 5.01%, 2.84% and 1.25% respectively. The average cocoon price of China in 2015 was forecasted to be 38.21 yuan/kg by using this method, increasing slightly compared to that in the last year. The forecast and verification results showed that this composite forecasting method has high accuracy and is suitable for forecasting cocoon price controlled by the market economic system.

关 键 词：蚕茧 价格预测 GM(1 1)灰色系统预测模型 BP人工神经网络模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] F323.7[自动化与计算机技术—控制科学与工程]