<![CDATA[Customer churn is a critical issue in the telecommunication industry as it directly affects a company's revenue. This study aims to develop and compare Logistic Regression and Support Vector Machine (SVM) models for predicting customer churn using the Telco Customer Churn dataset from IBM Watson Analytics, which consists of 7,043 customer records. The research process includes data exploration, data preprocessing, model training, and evaluation using Stratified K-Fold Cross-Validation (k = 5). The experimental results show that Logistic Regression achieved an accuracy of 80.70% with an average cross-validation score of 0.8043, while SVM achieved an accuracy of 79.28% with an average cross-validation score of 0.7954. Feature analysis indicates that tenure, MonthlyCharges, contract type, and internet service type are the most influential factors affecting customer churn. Based on these results, Logistic Regression demonstrates superior and more stable performance in predicting telecommunication customer churn.]]>
Copyrights © 2026
