{"title":"基于机器学习的方法预测保险公司的客户流失","authors":"Yunxuan He, Ying Xiong, Y. Tsai","doi":"10.1109/SIEDS49339.2020.9106691","DOIUrl":null,"url":null,"abstract":"Customer churn prediction plays an important role in business success for insurance companies like Markel Corporation. Each year Markel loses premium because some of their customers choose not to renew their policies. Based on the fact that the cost of attracting new customers is much greater than that of retaining existing customers, it is important for Markel to take early action to engage their customers before a policy expires. The goal in this work is to apply various machine learning methods and obtain an optimal model to predict customer churn rate. The dataset includes customer demographics features, customer behavior features, and macro environmental features. Exploratory analysis is conducted on critical features including policy length and types of coverage to draw insight about the impact of these features on the target variable – customers renew or do not renew their policies. With a large dataset, one of the main challenges is conducting feature dimension reduction and extracting important features to be used with a set of potential ML models. It turns out that the ML model with the best performance on the Area Under the Curve (AUC) metric is Extremely Randomized Trees Classifier and Gradient Boosting Model. Some suggestions on additional features to be incorporated are provided in the final comments. These features will improve predictive performance for the ML model of customer churn for Markel Corporation.","PeriodicalId":331495,"journal":{"name":"2020 Systems and Information Engineering Design Symposium (SIEDS)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2020-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"6","resultStr":"{\"title\":\"Machine Learning Based Approaches to Predict Customer Churn for an Insurance Company\",\"authors\":\"Yunxuan He, Ying Xiong, Y. Tsai\",\"doi\":\"10.1109/SIEDS49339.2020.9106691\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Customer churn prediction plays an important role in business success for insurance companies like Markel Corporation. Each year Markel loses premium because some of their customers choose not to renew their policies. Based on the fact that the cost of attracting new customers is much greater than that of retaining existing customers, it is important for Markel to take early action to engage their customers before a policy expires. The goal in this work is to apply various machine learning methods and obtain an optimal model to predict customer churn rate. The dataset includes customer demographics features, customer behavior features, and macro environmental features. Exploratory analysis is conducted on critical features including policy length and types of coverage to draw insight about the impact of these features on the target variable – customers renew or do not renew their policies. With a large dataset, one of the main challenges is conducting feature dimension reduction and extracting important features to be used with a set of potential ML models. It turns out that the ML model with the best performance on the Area Under the Curve (AUC) metric is Extremely Randomized Trees Classifier and Gradient Boosting Model. Some suggestions on additional features to be incorporated are provided in the final comments. These features will improve predictive performance for the ML model of customer churn for Markel Corporation.\",\"PeriodicalId\":331495,\"journal\":{\"name\":\"2020 Systems and Information Engineering Design Symposium (SIEDS)\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-04-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"6\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2020 Systems and Information Engineering Design Symposium (SIEDS)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/SIEDS49339.2020.9106691\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 Systems and Information Engineering Design Symposium (SIEDS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/SIEDS49339.2020.9106691","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Machine Learning Based Approaches to Predict Customer Churn for an Insurance Company
Customer churn prediction plays an important role in business success for insurance companies like Markel Corporation. Each year Markel loses premium because some of their customers choose not to renew their policies. Based on the fact that the cost of attracting new customers is much greater than that of retaining existing customers, it is important for Markel to take early action to engage their customers before a policy expires. The goal in this work is to apply various machine learning methods and obtain an optimal model to predict customer churn rate. The dataset includes customer demographics features, customer behavior features, and macro environmental features. Exploratory analysis is conducted on critical features including policy length and types of coverage to draw insight about the impact of these features on the target variable – customers renew or do not renew their policies. With a large dataset, one of the main challenges is conducting feature dimension reduction and extracting important features to be used with a set of potential ML models. It turns out that the ML model with the best performance on the Area Under the Curve (AUC) metric is Extremely Randomized Trees Classifier and Gradient Boosting Model. Some suggestions on additional features to be incorporated are provided in the final comments. These features will improve predictive performance for the ML model of customer churn for Markel Corporation.