TELECOMMUNICATIONS SUBSCRIBER CHURN CLASSIFICATION USING SERVICE USAGE AND BILLING BEHAVIOR

Abstract

In the highly competitive telecommunications industry, retaining existing customers is as
important as acquiring new ones. Customer churn, which refers to subscribers discontinuing
their services, poses a significant challenge for telecom companies. Predicting churn in
advance helps organizations take proactive measures to improve customer satisfaction and
reduce revenue loss. This project focuses on developing a machine learning model to classify
telecommunications subscribers based on their likelihood of churn using service usage
patterns and billing behavior.
The dataset includes customer information such as call duration, internet usage, billing
amounts, payment methods, contract type, and customer tenure. These features provide
insights into customer engagement and financial interactions with the service provider.
Various data preprocessing techniques such as data cleaning, feature encoding, and
normalization are applied to prepare the dataset for model training. Machine learning
algorithms such as Random Forest, Logistic Regression, or Decision Tree are used to build
the churn prediction model.
The model analyzes customer behavior patterns and identifies subscribers who are at risk of
leaving the service. The results demonstrate that analyzing service usage and billing behavior
can effectively predict customer churn. This predictive approach enables telecom companies
to implement targeted retention strategies, improve customer satisfaction, and enhance
business profitability. The proposed system provides a practical solution for churn
management using data-driven decision-making.