Использование прогнозной аналитики и анализа неструктурированных данных для предотвращения эскалации проблем и оттока клиентов

Abstract

This report is aimed at describing a practical approach to building a predictive model that helps establishing long-term relations with your customers by increasing their loyalty and reducing churn. The model allows responding to early incidents by proactively leveraging additional resources to solve potential problems before they actually lead to critical situations. The model is based on mathematically grounded forecasts, which are built by analyzing structured and unstructured incident data using IBM analytical software platforms.