基于决策树的客户流失预测模型Customer loss prediction model based on decision tree
张静怡;胡俊英;李卫斌;
摘要(Abstract):
随着证券市场竞争的日益加剧,证券行业客户数量呈现动态增长模式,但是在大量客户开户的同时,又有大批客户流失,带来较多的无交易客户,导致业务与收入总量增长相对趋缓,出现“增量不增收”的现象.准确预测潜在流失客户,对这些客户实施差异化营销和服务已成为当前证券企业的迫切需求.基于证券客户交易的历史数据,在给出证券流失客户定义的基础上,选择合适的自变量和因变量时间窗口,建立信息熵,趋势值和波动值三类特征指标体系,使用CART算法建立决策树模型,并采用交叉验证法选取最优决策树,实现依据历史交易数据对客户是否流失进行预测.数据实验结果表明,所提出的预测方法可以较准确地对潜在流失客户进行预测.
关键词(KeyWords): 客户流失;衍生指标;决策树;CART算法;预测模型
基金项目(Foundation): 国家自然科学基金(12001428)
作者(Authors): 张静怡;胡俊英;李卫斌;
参考文献(References):
- [1]肖进,刘敦虎,贺昌政.基于GMDH的“一步式”客户流失预测集成建模[J].系统工程理论与实践,2012,32(4):807-814.
- [2]胡永培,张琛.基于AP聚类与随机森林的客户流失预测研究[J].计算机技术与发展, 2021,31(2):49-53.
- [3]周静,周小宇,王汉生.自我网络特征对电信客户流失的影响[J].管理科学, 2017,30(5):28-37.
- [4]王伟钧.基于数据挖掘的证券营业部客户流失分析[J].电子科技大学学报(社会科学版), 2009,11(1):18-22.
- [5] Pan W W. Fraudulent firm classification using monotonic classification techniques[C]//Proceedings of the IEEE 9th Joint International Information Technology and Artificial Intelligence Conference.Piscataway:IEEE, 2020:1773-1776.
- [6] Erin M, Vidita G, Weihong G G. A physics-informed convolutional neural network with custom loss functions for porosity prediction in laser metal deposition[J]. Sensors, 2022,22(2):494-512.
- [7] Hong L P, Qin X Zh, Jia Zh H, et al. Prediction support vector machine based on posterior probability in loss of customers[J]. Computer Engineering and Design, 2016,37(2):429-432.
- [8] Cano J R, Guti′errez P A, Krawczyk B, et al. Monotonic classification:an overview on algorithms,performance measures and data sets[J]. Neuro Computing, 2019,341(1):168-182.
- [9] Zeng R, Yuan L, Ye Z, et al. Prediction and analysis model of telecom customer churn based on missing data[C]//Conference on Advanced Computer Architecture. Singapore:Springer, 2020:221-232.
- [10]李爱民.基于逻辑回归和聚类分析的广电客户流失预警[J].信息与电脑(理论版), 2016,9(1):29-30.
- [11] Xia S Y, Wang G Y, Chen Z Z, et al. Complete random forest based class noise filtering learning for improving the generalizability of classifiers[J]. IEEE Transactions on Knowledge and Data Engineering, 2019,31(11):2063-2078.
- [12]林明辉.基于BP网络的通讯行业客户流失预警模型研究[J].陕西学前师范学院学报, 2016,32(3):146-149.
- [13] Swetha P, Dayananda B. Improvised XgBoost machine learning algorithm for customer churn prediction[J]. EAI Endorsed Transactions on Energy Web, 2020,7(30):51-54.
- [14]张世霖. CART分层变量的选择[D].成都:电子科技大学, 2019.
- [15]张俊玉,胡家豪,黄嵩. CART决策树方法在煤电厂节能降耗中的应用[J].控制与决策, 2021,36(5):1232-1238.