Churn Prediction Application in The Business World
Grade Level at Time of Presentation
Senior
Major
Mathematical Economics
Minor
Applied Statistics
Institution
Western Kentucky University
KY House District #
17
KY Senate District #
32
Faculty Advisor/ Mentor
Dr. Lily Zhuhadar
Department
Business Data Analytics
Abstract
The latest escalation in the trade war between the USA and China is bad news for stakeholders, financiers, and investors. Businesses are becoming worried that a significant market crash is imminent. Within this context, churn prediction is considered one of the most critical techniques in Business-to-Business (B2B) contexts. In this research study, we focus on churn prediction modeling. When considering B2B churn prediction, the research that can be found is limited. To address this gap in research, we perform a benchmarking study of churn prediction techniques in a B2B context. The predictive power of logistic regression, decision trees, random forest, neural networks, support vector machines are evaluated on a dataset.
Regarding performance, accuracy, AUC, sensitivity, specificity, F-measure, and top decile lift are calculated. When considering variable importance, we identify the essential recency and frequency variables for every prediction technique. Furthermore, the findings suggest that the significance of specific categories of variables may vary depending on the applied prediction technique. To summarize, the contribution of this study is an analysis of classification techniques that have been formerly used in B2B and B2C churn prediction is presented and evaluated.
Churn Prediction Application in The Business World
The latest escalation in the trade war between the USA and China is bad news for stakeholders, financiers, and investors. Businesses are becoming worried that a significant market crash is imminent. Within this context, churn prediction is considered one of the most critical techniques in Business-to-Business (B2B) contexts. In this research study, we focus on churn prediction modeling. When considering B2B churn prediction, the research that can be found is limited. To address this gap in research, we perform a benchmarking study of churn prediction techniques in a B2B context. The predictive power of logistic regression, decision trees, random forest, neural networks, support vector machines are evaluated on a dataset.
Regarding performance, accuracy, AUC, sensitivity, specificity, F-measure, and top decile lift are calculated. When considering variable importance, we identify the essential recency and frequency variables for every prediction technique. Furthermore, the findings suggest that the significance of specific categories of variables may vary depending on the applied prediction technique. To summarize, the contribution of this study is an analysis of classification techniques that have been formerly used in B2B and B2C churn prediction is presented and evaluated.