作　　者： ;

机构地区： 广州南洋理工职业学院

出　　处： 《北京印刷学院学报》 2023年第9期21-25,共5页

摘　　要： 本文采用Adaboost算法对库存成本完成多元分类,通过Adaboost多个弱分类器的加权作用,获得不同特征样本下的库存成本预测分类结果,为库存管理提供策略支持。实验结果表明,合理设置弱分类器数量及学习率,集成学习Adaboost算法可以获得较高的库存管理经济预测精度,在3个不同行业的库存管理实例应用中均取得了较高的分类预测性能,这说明集成学习算法在库存管理经济预测的适应度高。 In this paper,Adaboost algorithm is used to complete multivariate classification of inventory cost.By weighting multiple weak classifiers of Adaboost,inventory cost prediction classification results under different feature samples are obtained,which provides strategic support for inventory management.The experimental results show that the integrated learning Adaboost algorithm can obtain higher accuracy of inventory management economic prediction with reasonable setting of the number and learning rate of weak classifiers,and has obtained higher classification prediction performance in three different inventory management examples,which indicates that the integrated learning algorithm has high adaptability in inventory management economic prediction.

关 键 词： 库存管理 集成学习 算法 成本优化

领　　域： [自动化与计算机技术—计算机系统结构] [自动化与计算机技术—计算机科学与技术]