Customer Churn Prediction Using Machine Learning Techniques: In the case of Awash Bank Wolaita Sodo Region

Abstract

Churn prediction refers to the procedure of identifying customers who are highly likely to terminate their service subscription based on their utilization. Since there is still time to take preventative action, being able to predict a client who is likely to leave is essential for solving business problems. The banking industry in Ethiopia currently has millions of users, making it challenging to analyze and anticipate consumer attrition. There is diverse research that has been conducted in this particular domain. The primary challenges encountered in the majority of the prior investigations were associated with the selection of the suitable technique for achieving data balancing, the predicaments revolving around the choice of a technique for handling missing values, and the excessive dependence of the model on a singular attribute. The aim of this research is to develop a machine-learning model that can predict customer churn for the Awash Bank Wolaita Sodo region. The dataset utilized for this investigation comprises 50,987 entries encompassing 11 attributes, which were collected from this region. Among these, 31,619 represent active accounts, while the remaining 19,368 pertain to closed accounts. To achieve balance within the dataset, a hybrid method is employed, while an extraction tree classifier is employed for the purpose of selecting features. EdrawMax was used for creating diagrams, and Python was used for implementation. This research used an experimental research approach, and eight algorithms are used, such as extreme gradient boosting (XGBoost), random forest, light gradient boosting machine (LGBM), decision tree, one-dimensional convolutional neural network (1D-CNN), gradient boosting machine (GBM), deep neural network (DNN), and multi-layer perceptron (MLP). For evaluating the performance of the model, accuracy, f1-score, recall, and precision are used. Based on the results of the models, the random forest model outperformed other models with an accuracy of 99.14% and 99% recall, precision, and f1-score. Based on the findings of this study, it is highly recommended that the feature researcher undertakes further research for Awash Bank in various regions. This can be achieved by increasing the dataset and extending the scope of research to other banking industries