Artificial Intelligence Based Customer Churn Prediction Model for Business Markets

Area of Science:

• Artificial Intelligence
• Machine Learning
• Telecommunications

Background:

• Customer churn significantly impacts telecommunication company revenue.
• Developing effective Customer Churn Prediction (CCP) models is crucial for the telecom industry.
• Existing AI and ML models show promise for CCP solutions.

Purpose of the Study:

• To propose a unique AI-based CCP model for Telecommunication Business Markets (AICCP-TBM).
• To control the identification of churners and non-churners in the telecom sector.
• To enhance classification performance for improved customer retention strategies.

Main Methods:

• Utilizing Chaotic Salp Swarm Optimization-based Feature Selection (CSSO-FS) for optimal feature selection.
• Employing a Fuzzy Rule-based Classifier (FRC) to differentiate between churners and non-churners.
• Optimizing FRC membership functions using Quantum Behaved Particle Swarm Optimization (QPSO).

Main Results:

• The AICCP-TBM model demonstrated superior performance compared to state-of-the-art CCP models.
• Achieved high accuracy rates of 97.25%, 97.5%, and 94.33% on three benchmark datasets.
• Validated improved prediction performance through extensive experimental analysis.

Conclusions:

• The proposed AICCP-TBM model offers an effective solution for customer churn prediction in telecommunications.
• The integration of CSSO-FS and QPSO-optimized FRC enhances predictive accuracy.
• This AI-driven approach aids telecom businesses in proactive customer retention efforts.