基于客户画像和GBDT算法的客户价值预测方法

Customer Value Prediction Method based on Customer Portrait and GBDT Algorithm

针对现今客户价值的差异对企业服务运营的影响,本文基于对Superstore Sales客户消费数据的分析,提取客户特征构建基于消费行为的客户画像;抽取客户价值评价指标,经主成分分析法(PCA)优化后形成客户价值评价指标体系;利用K-Means聚类算法为客户贴上价值标签,构建基于梯度提升决策树算法(GBDT)的客户价值预测模型。实验结果表明,该方法预测结果具有较高的准确率,可挖掘隐藏的高价值客户,对于企业客户关系管理具有一定的研究借鉴价值与实际意义。

In view of the impact of the difference in customer value on the service operation of the enterprise, this paper is based on the analysis of customer data of Superstore Sales, extracts customer characteristics and builds customer portraits based on consumer behavior; then evaluation index system is formed via abstracted evaluative criteria of customer value through principal component analysis （PCA） ; followed by building customer value prediction model based on gradient boosted decision tree （GBDT） algorithm via classifying customers by their value using K - Means clustering algorithm. The experimental results show that the method has a high accuracy in predicting the results and can be used to mine hidden high value customers. It has certain reference value and practical significance for customer relationship management.