基于层级提升重校正模型的备件分类研究-《计算机技术与发展》

文章信息/Info

Title:: Research on Classification Strategy of Spare Parts Based on Cascade Boosting and Recalibrating Model

Author(s):: WANG Hai-yang; WU Qi-shi; Southwest Jiaotong University,Chengdu 611756,China

Keywords:: auto parts; parts classification; cascade boosting; ensemble model; deep forest

摘要:: 车辆备件销售占汽车产业链整体价值的比重逐年增大,并且已经成为了部分汽车企业的主要营收来源。在售后服务中车企是否能及时供应车辆备件更是成为凸显其服务水平高低和品牌价值的重要指标。而车辆备件库存受时间和市场等外部因素的影响较大,如何合理地将合适的管控策略应用到正确的备件上已成为降低库存风险和提升企业售后服务水平的一个难题。文章从车辆备件的多类别属性出发,并以某真实汽车企业为研究对象,提取出了其在全国具有代表性的售后服务商的配件库存特征,并在此数据基础上,提出了一种基于层级提升重校正的集成分类模型(CBRE)用于合理地分类备件,为其分配合适地管控策略。此模型相较于其他分类模型具备分类精度高且能够根据实际数据自适应调整模型结构复杂度的特点。

Abstract:: The proportion of auto parts sales in the overall value of the automobile production chain is increasing year by year, and has become the main revenue source for some automobile enterprises. Whether automobile enterprises can supply auto parts in time in aftersales service has become? ? an important indicator to highlight their service level and brand value.The inventory of auto parts is greatly influenced by external factors such as? ?time and market. How to reasonably apply the appropriate control strategy to the right auto parts has become a difficult problem to reduce inventory risk and improve the after-sales service level of enterprises. Starting from multi-category attributes of auto parts,taking a real automobile enterprise? as the research object,extracting the auto parts features of the representative after-sales service providers,and on the basis of this data,we propose an auto parts classification model based on cascade boosting and recalibrating ensemble model to classify auto parts reasonably and allocate appropriate control strategy for them. Compared with other classification models,the proposed model has the characteristics of high classification accuracy and can adjust the structure complexity of the model adaptively according to the actual data.