Modeling the Customer Time to Churn-The Application of the Logistic Regression on the Private Security Company

• Main Authors: CHANG, HSUAN-YEH, 張玄燁
• Other Authors: LIANG, TE-HSIN
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/50338344728580917182

碩士 === 輔仁大學 === 統計資訊學系應用統計碩士班 === 105 === This study mainly discuss factors that cause the customer churn in the private security company and use the Logistic Regression to create the predict model. The study use the downgrade of the customer to be the dependent variable, and make the independent variable to three classifications: The property of the customer, the behavior of the customer and the additional data, the study choose the population rate of crime to be the additional data. The study use The Chi-square Test and The Correlation Analysis to discuss the relation between factors and use factors to make five independent Logistic Regression Models: Taipei city, New Taipei City, Taichung City, Tainan City and Kaohsiung City. Our data analysis shows that the property of the customer is the common factor of five models that affects the customer churn, especially the business type that customer worked and Number of years in service of private security company . The behavior of the customer affects the customer churn in Taipei City New Taipei city, Tainan City and Kaohsiung city, it’s not affects the Taichung City. The additional data doesn’t affect models of five cities.