The goal of this project is to develop a predictive model to identify customers likely to churn within a telecommunications company. Anticipating customer churn can significantly benefit the company by reducing the costs associated with customer acquisition.
We employed supervised machine learning techniques to build models capable of classifying customers as potential churners or non-churners based on their historical usage patterns and service plan features. Our analysis included popular algorithms such as Logistic Regression, Support Vector Machines (SVM), Decision Trees, K-Nearest Neighbors (K-NN), and XGBoost.
After thorough evaluation, the Decision Tree model emerged as the top performer, achieving a balanced accuracy of 97.86%. Although the ensemble method with XGBoost slightly outperformed it with a balanced accuracy of 98.06%, we opted for the Decision Tree due to its favorable balance between performance and computational efficiency.
The chosen Decision Tree model serves as a valuable tool for the company, enabling proactive identification of customers at risk of churn. This allows timely implementation of retention strategies to mitigate churn rates effectively.
Fahad Alsubaie
Haaniya Umair
Jiayi Huang
Parita Patel
Yifei Cheng