时间序列和神经网络对迷你型洗衣机销量预测  

Mini Washing Machine Sales Prediction Based on Time Series and Neural Networks

作　　者：石磊[1] 

机构地区：[1]阜阳职业技术学院信息技术系,安徽阜阳236031

出　　处：《安徽理工大学学报（自然科学版）》2013年第3期69-73,共5页Journal of Anhui University of Science and Technology:Natural Science

基　　金：安徽省自然科学研究资助项目(KJ2011B129)

摘　　要：迷你型洗衣机价格便宜,使用方便,近年来销售量不断增加。缺点是功能单一导致消费受众面窄,同时受季节性因素影响较大,企业往往难以准确制定生产销售计划,导致库存和缺货现象时有发生。将合作企业生产的迷你型洗衣机近4年销售数据和影响销售的关联因素作为训练样本,先建立时间序列回归模型和BP神经网络模型对洗衣机销量进行预测,然后建立遗传算法优化的灰色神经预测模型。通过对三种模型预测结果的对比分析,表明经由遗传算法优化参数的灰色神经网络能较好的预测销量未来变化,辅助企业决策订单生产计划。Mini washing machines are cheap and easy to use , their sales have been increasing for recent years . The disadvantage of simple function results in narrow range of users , and their sales are heavily affected by sea-sons.So it’s hard for enterprises to set production and sales plan accurately , which leads to inventory and short-age occrence .The sales data and associated factors affecting sales of mini washing machine produced by coopera -tive enterprises of recent four years were used as training samples to establish time series regression model and BP neural network model to predict washing machine sales and then the gray neural prediction model optimized by genetic algorithm was built .Comparative analysis of the prediction results with these three models showed that prediction of sales changes in the future by the gray neural prediction model optimized by genetic algorithm is better , which asists the enterprises to make production plan for ordering .

关 键 词：订单预测 灰色神经网络 遗传算法 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]