Machine Learning-Based Customer Churn Prediction for E-Commerce Businesses

The problem of customer churn remains a major concern in the e-commerce industry, as it directly impacts a company's revenue and long-term growth. This study employs a machine learning approach to develop predictive models using a dataset comprising 5,630 samples and 20 variables. Missing values were handled, 48 outliers were removed, and forward feature selection was applied for analysis and model training. This research focuses on supervised machine learning, utilizing classifiers such as Random Forest, Gradient Boosted Trees, k-Nearest Neighbors, and Decision Trees. Among them, Random Forest demonstrated the best performance in terms of accuracy, precision, and recall. Hence, this study aims to implement advanced machine learning algorithms to help e-commerce businesses build robust churn prediction models and improve customer retention.