Machine Learning Based Customer Churn Prediction on Telecom Industry

Given the expenses incurred, customer churn stands as a significant worry for telecom service providers. The financial implications linked with customer churn make it a pressing concern for these providers. This brief article provides insight into our ongoing research efforts focused on forecasting customer attrition within the telecommunications service sector. Anticipating customer churn remains a contemporary challenge in the telecommunications industry, crucial for comprehending customer patterns and pinpointing those inclined to end their subscriptions. Customer churn, the percentage of customers discontinuing services, is tackled using Random Forest Classifiers and Support Vector Machine models in our study. Through grid search, we assess their effectiveness for predicting telecom customer turnover. This research applies ml and dl to proactively manage retention, considering the greater costs of getting new people(customer) compared to keep hold of existing ones. The goal is to equip businesses with predictive tools for efficient resource allocation and targeted retention strategies. From the results of our experiment, we can conclude that the Random Forest model exhibits superior performance for this specific use case, boasting a prediction accuracy of 95.96% on the examining data. This emphasizes the effectiveness of Random Forest model in tackling the crucial problem of churn in the telecom sector.