Akademik Veri Yönetim Sistemi
Economics World, cilt.12, sa.2, ss.142-154, 2025 (Hakemli Dergi)
This study evaluates the use of predictive analytics to forecast customer turnover in subscription-based Services in order to develop a predictive model to help small and medium-sized enterprises manage customer churn in the face of digital disruption. The research uses a quantitative approach focusing on empirical customer data to accurately predict buying trends and adapt marketing techniques. Demand forecasts in the health sector are important, as in every sector. In particular, the material forecast and stock forecasting of the purchasing unit of hospitals are among the areas that receive significant attention. Four classifiers (Random Forest, Logistic Regression, Gradient Boosting and XGBoost) are trained and evaluated using various performance indicators as part of a systematic approach involving Kaggle data collection, preparation and model selection. The results show excellent accuracy in predicting customer attrition, but there are limitations in precision and recall, indicating room for improvement. Confusion matrices provide information about the performance of each classifier, allowing for continuous improvement of predictive analytics techniques. Ethical concerns are rigorously addressed throughout the work process to guarantee appropriate data and machine learning methodologies. The proposals emphasize the proactive use of predictive analytics to identify at-risk customers and implement targeted retention strategies. Incorporating new data sources, improving customer experience, and utilizing collaborative churn management methods are recommended to increase forecast accuracy and business outcomes. Finally, this research provides important insights into the usefulness of predictive analytics for customer churn forecasting as well as practical recommendations for businesses seeking to increase customer retention and reduce churn risk. By leveraging empirical research findings and implementing ethical and rigorous churn control strategies, businesses can achieve long-term success in today’s changing market environment. Keywords: artificial intelligence, customer behavior, health sector, prediction, analytics