A Comparison of Feature Selection Algorithms with Explainable AI

Sanluang, Pattamaporn (2024) A Comparison of Feature Selection Algorithms with Explainable AI. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (4MB) | Preview

Abstract

In the telecommunication sector, losing a customer has a direct negative impact on revenue and long-term growth prospect, making customer churn prediction a major business priority. To develop successful retention strategies, an understanding of the causes of customer churn is required. However, traditional black-box models often lack the transparency needed to interpret insights into actionable decisions. This research employed Explainable AI (XAI) techniques with classification models, namely Feature Importance, SHAP, and LIME to select the top features influencing churn. This approach enhances the interpretability of churn prediction models without compromising accuracy. Results show that XAI-based feature selection models were easier to interpret (smaller models), are easier to interpret from business point of view, and were able to obtain a higher true positive rate of customer churn. While the baseline Random Forest model achieved the highest F1-score of 79% with all the features, a subset of features selected using SHAP only decreased F1-score to 77%, yet still provided valuable insights into feature contribution. Selected features were able to correctly detect a higher true positive, which in this context is the correct customers that will leave the company. In other words, the confusion matrix revealed that the retrained XGBoost model produced more true positive predictions, which confirms the benefits of targeted feature selection. Demonstrates the potential of XAI to balance model accuracy with transparency offering telecommunications companies a more informed approach to customer retention strategies.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Estrada, Giovani UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence H Social Sciences > HF Commerce > Marketing > Consumer Behaviour Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Telecommunications Industry
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	25 Aug 2025 10:59
Last Modified:	25 Aug 2025 10:59
URI:	https://norma.ncirl.ie/id/eprint/8623

Actions (login required)

View Item