魏贝贝 卢明银

摘 要： 空调行业一直处于双寡头格局，从双十一的价格战到目前的疫情影响，中小企业的电子商务销量预测面临着更加严峻的考验。为降低传统预测模型的低拟合度，提高预测精度，提出了布谷鸟算法与BP神经网络算法相结合的预测方法。首先对影响空调销量预测的主要因素进行了降维处理，其次运用布谷鸟算法对BP神经网络算法的权值和阈值进行优化，同时在布谷鸟算法中引入了细菌觅食算法的群体信息共享机制，加快算法的收敛。通过与不同预测方法的结果对比，该方法的预测误差更低，可以为空调电子商务销量预测提供良好的科学依据。

关键词： 因子分析;布谷鸟算法;BP神经网络算法;销量预测

中图分类号： TT-9

文献标志码： A

Air Conditioning E-commerce Based on ImprovedCuckoo-BP Algorithm Sales Forecast

WEI Beibei LU Mingyin

（1.School of Mining Engineering， China University of Mining and Technology，Jiangsu Xuzhou 221100， China; 2.School of Mining Engineering， China Universityof Mining and Technology， Jiangsu Xuzhou 221100， China）

Abstract： The air-conditioning industry has always been in a duopolistic pattern. From the price war of the Double Eleven to the current epidemic， the forecast of SMEse-commerce sales is facing a more severe test. In order to reduce the artificial interference and the low fitting degree of the traditional prediction model and improve the prediction accuracy， it will propose a prediction method combining the cuckoo algorithm and the BP neural network algorithm. First of all， the main factors that affect the sales forecast are processed for dimensionality reduction. Secondly， the cuckoo algorithm is used to optimize the weights and thresholds of the BP neural network algorithm. At the same time， the cuckoo algorithm introduces the group information sharing mechanism of the bacterial foraging algorithm to speed up the convergence of the algorithm. Compared with the results of different prediction methods， the prediction error of this method is lower， which can provide a good scientific basis for the sales forecast of e-commerce of air conditioners.

Key words： factor analysis; cuckoo algorithm; BP neural network algorithm; sales forecast

2019年網购零售额达106324亿元，比上年增长16.5%，网络零售在促进消费结构升级、拉动内需等方面起着越来越重要的作用，京东、苏宁易购、天猫等电商平台成为了推动消费升级的主力军。与此同时，企业运营问题面临着更大的挑战。由于客户在多个平台的选择性广，且企业无法直接接触客户，特别是B2C模式更增加了订单预测的难度。如果生产过多，引起库存积压，占用大量财务资金影响企业正常运转，生产过少，缺货状态下只能形成越卖越差的恶性循环，而且一旦预测不准确，频繁的变更订单需求，势必会影响采购、生产环节的工作安排，导致工厂间产能、物料的不均衡，因此保证订单预测高准确率和低波动率显得尤为重要。……