Explainable AI-Based Deep Learning System for Predicting Customer Churn in Telecommunication Industry

An Explainable Machine Learning Approach With Density-Regularized Oversampling for Industrial Customer Churn Prediction

Customer churn prediction has become a vital statistical endeavour for businesses seeking to keep clients and increase their profitability over time. Although numerous data-driven models have been proposed, churn prediction remains difficult due to complex customer behaviours, heterogeneous attributes, and high overlap between churners and non-churners. Existing studies proposed various machine learning (ML) and deep learning (DL) models along with improved feature engineering. However, the fundamental research gap is a lack of attention to class imbalance handling techniques that retain inherent data distribution in churn prediction tasks. More specifically, conventional oversampling approaches frequently generate synthetic instances without taking into consideration local data density, resulting in noisy samples and poor classifier performance. This research suggests a density-regularized SMOTE-ENC framework that specifically incorporates neighbourhood-based density restrictions into the process of creating synthetic samples for mixed-type churn datasets. The proposed method selectively retains synthetic minority samples that lie closer to minority class manifolds while being sufficiently distant from majority class regions. The proposed density-regularized oversampling framework consistently improves churn prediction performance across all datasets, achieving 84.00% accuracy, 60.16% precision, 64.46% recall, and 62.14% F1-score on Bank Churn; 77.44% accuracy, 55.38% precision, 77.21% recall, and 64.49% F1-score on Telco-IBM; and 95.65% accuracy, 87.46% precision, 87.24% recall, and 87.09% F1-score on Retail Churn. Compared with the original imbalanced datasets without oversampling, the proposed method improves the F1-score by 7.3%, 3.1%, and 0.5% on the Bank, Telco-IBM, and Retail datasets, respectively.

Customer Churn Prediction in the Telecommunications Sector Using Explainable Artificial Intelligence

: Customer churn poses a significant threat to profitability in the saturated telecommunications industry, yet accurately predicting churn remains challenging due to high-dimensional data and class imbalance where churners represent a small minority. While complex ensemble methods achieve high accuracy, their ”black-box” nature limits business adoption, as practitioners require transparent insights to design effective retention campaigns. This paper proposes a comprehensive Machine Learning pipeline that bridges the gap between predictive performance and model interpretability. We implement and compare five classifiers—Logistic Regression, Random Forest, XGBoost, LightGBM, and CatBoost—optimized via Bayesian hyperparameter tuning and evaluated using the recall-focused F2-score to address class imbalance. Our results demonstrate that gradient boosting models, particularly XGBoost, outperform aggregation-based ensemble strategies, achieving the highest F2-score of 0.7500 with a recall of 0.9385. Crucially, we integrate SHAP-based Explainable Artificial Intelligence to provide both global and local interpretability, revealing that contract type, tenure, and technical support subscriptions are the primary churn drivers. This dual-layer transparency enables targeted retention strategies while preserving high recall in imbalanced telecom data.

A Binary Firefly Optimized Stacking Ensemble Model for Customer Churn Prediction in the Telecommunications Industry

Customer churn prediction remains a critical challenge in the telecommunications industry because subscriber attrition directly affects revenue generation, customer retention, and long-term business sustainability. Although numerous studies have employed machine learning and deep learning techniques for churn prediction, many existing approaches rely on standalone models that have limited ability to simultaneously capture complex nonlinear customer behaviour, exploit model diversity, and eliminate redundant features. Building upon our previous study, this research proposes an enhanced heterogeneous stacking ensemble framework that integrates Binary Firefly Algorithm (BFA)-based feature optimization with Deep Neural Network (DNN), Support Vector Machine (SVM), and Random Forest (RF) as base learners, while Logistic Regression serves as the meta-learner for final churn classification. To ensure a fair and controlled comparison, the same Maven Analytics Telecom Customer Churn dataset and preprocessing strategy adopted in the previous study were retained, including data cleaning, feature transformation, stratified data partitioning, and normalization. Model development further incorporated 5-fold cross-validation on the training dataset, while BFA was introduced to identify the most informative pre-churn features. Experimental results demonstrate that the proposed framework achieved strong and balanced predictive performance and demonstrated improvements in several classification metrics compared with the previously developed BFA-based Hybrid TabNet-DNN model. Furthermore, its performance was competitive with the Random Forest baseline, which exhibited comparable classification effectiveness. These findings show that optimized feature selection and heterogeneous ensemble learning improve prediction stability and generalization. The proposed framework provides an effective decision-support tool for proactive customer retention in the telecommunications industry.

XGBoost Algorithm for Telecom Customer Churn Prediction and Its Business Implications

Predicting customer loss is a critical challenge for telecommunications companies, as identifying at-risk customers is key to reducing financial losses and formulating effective retention strategies. Conventional statistical methods and basic ensemble models often perform poorly on telecom datasets due to their inability to capture complex feature relationships and handle class imbalance. This study applies the Extreme Gradient Boosting (XGBoost) algorithm to telecom customer churn prediction, comparing its performance with Logistic Regression and Random Forest using the public Telco Customer Churn dataset (7,043 records, 21 features). A rigorous data preprocessing pipeline was implemented, including missing value handling, categorical encoding, and feature standardization, with an 80/20 train-test split that preserved class distribution. Hyperparameter tuning for XGBoost addressed the 26.5% churn rate imbalance. Evaluation metrics (accuracy, precision, recall, F1- score, AUC) showed XGBoost outperformed benchmark models: 82.1% accuracy, 78.3% churn recall, 59.2% precision, 67.4% F1-score, and 0.869 AUC. Feature importance analysis identified customer tenure, contract type, monthly charges, and fiber optic internet service as the primary churn drivers. XGBoost also captured complex non-linear interactions unrecognized by other models.

Hybrid Deep Neural Network with Natural Language Processing Techniques to Analyze Customer Satisfaction with Delivery Platform Manager Responses

Predicting Telecom Customer Churn and Win‑Back Probability: A Stacked Ensemble Approach