摘要
需求预测对于宏观和微观经济研究具有重要意义,其分析方法有很多种,灰色神经网络是其中一种。但是,灰色神经网络需要进行修正才能保证预测的稳定性。为解决这一问题,将人工蜂群算法引入灰色神经网络的改进中,并将其应用于冰箱订单需求预测中。采用人工蜂群算法优化基本灰色神经网络的白化参数,得到最优个体后,将其值作为灰色神经网络的初始权值和阈值。对订单需求的预测实验结果表明,相对于基本以及其它优化的灰色神经网络模型,采用人工蜂群改进的灰色神经网络模型预测误差更小、精度更高,为优化灰色神经网络提供了一个新方法,也拓展了需求预测的新思路。
Demand forecasting is of great significance to macro and micro economic research. There are many kinds of analysis methods, and grey neural network is one of them. However, the grey neural network needs to be modi-fied to ensure the stability of the prediction. In order to solve this problem, the artificial bee colony algorithm is introduced into the improvement of the grey neural network. The artificial bee colony algorithm is used to optimize the whitening parameters of the basic grey neural network, and the optimal individual is obtained. The experimental resuits show that, compared with the basic and other optimized grey neural network model, the improved grey neural network model based on artificial bee colony has smaller prediction error and higher precision. It provides a new method to optimize the grey neural network, and expands the new idea of demand forecast.
出处
《宜春学院学报》
2017年第6期46-49,共4页
Journal of Yichun University
基金
安徽省哲学社科规划项目(AHSKY2015D71)
安徽省社科创新发展研究课题(A2015020)
关键词
人工蜂群
灰色神经网络
需求预测
artificial bee colony
grey neural network
demand forecasting
