Extracting Appropriate Features for Behavior Analysis and Predicting Customer Attrition in Banks

Author(s):
Message:
Article Type:
Research/Original Article (بدون رتبه معتبر)
Abstract:

Customer loyalty has always been one of the main concerns and challenges for domestic banks, as their lifecycle depends on customer loyalty. Various studies have been conducted on customer loyalty in banks and financial and commercial institutions. This research evaluated the use of data mining and five machine learning algorithms—Support Vector Machine, Naive Bayes, MLP, Decision Tree, and Logistic Regression—for detecting customer loyalty in a domestic bank. Three performance metrics—accuracy, precision, and recall—were used to compare the results. The data used in this research were collected from a domestic bank and consisted of 15 criteria for evaluating customer loyalty to the organization. The results showed that, among the five algorithms, the best performance was achieved by the Decision Tree, while the Logistic Regression algorithm showed the worst performance. Finally, to improve the results, a combination of a neural network and the Teacher-Student optimization algorithm was used to achieve better detection of loyal customers.

Language:
Persian
Published:
Journal of New Researches in the Smart City, Volume:3 Issue: 1, 2025
Pages:
29 to 49
https://www.magiran.com/p2822505