中文摘要：

电子产品型号类型多、更新速度快等特点导致了预测回收量的难度大,因此对第三方维修服务商而言,电子产品回收量预测的精度直接影响到企业的运营成本以及服务水平.通过企业真实数据的回归分析,发现产品的累计销售量与累计回收量之间存在显著的线性相关性,由此设计了回归预测方法与阻尼趋势预测方法相结合的组合预测方法,并进行数值实验.实验结果表明该组合预测方法在电子产品回收预测量中能达到比使用单个模型更好的效果,实现了预测精度的显著提升.

英文摘要：

Characteristics of electronic products such as multiplicity and rapid renewal render returns more difficult to forecast.For third party maintenance service providers,the accuracy of electronic products returns forecasting plays a crucial role in terms of operation cost and service level.This paper uses a company＇s real data to find that there exits a linear relationship between its cumulative sales and cumulative returns quantity by linear regression,and develops a new forecasting method combining regression forecasting method with damped trend exponential smoothing method.Results of the numerical experiment show that this combination method performs better in practice than those single models do,and the forecasting accuracy can be improved significantly.