关键词: 汽车售后服务, 客户细分, RFMD模型, 半监督谱聚类集成（SSSCE）算法, 保养策略

Abstract: According to the purpose of automotive after-sales service customer segmentation and the maintenance of the vehicle during the warranty period, the RFMD customer segmentation indicator model is constructed. Aiming at the advantages that the clustering ensemble algorithm can fully exploit the internal structure of the data set, and the semi-supervised learning idea uses the prior knowledge to guide the clustering, the Semi-Supervised Spectral Clustering Ensemble（SSSCE） algorithm is applied to after-sales customer segmentation. Compared with the Spectral Clustering（SC） algorithm and the Spectral Clustering Ensemble（SCE） algorithm, the SSSCE algorithm has better customer segmentation results. Finally, the customer group segmented by SSSCE algorithm is analyzed for characteristics, and the corresponding maintenance guidance strategy is given.

Key words: automotive after-sales service, customer segmentation, RFMD model, Semi-Supervised Spectral Clustering Ensemble（SSSCE） algorithm, maintenance strategy