Prediction of Financial Distress in Bangladesh’s Banking Sector Using Data Mining and Machine-Learning Technique

Abstract

Now bank failures in a developing country like Bangladesh are a prevalent and delicate problem. It is therefore essential to examine and predict a bank’s economic health so that it can assist to minimize and rectify the bank’s and customer’s future or present losses. This research work therefore focuses on using distinct approaches to machine learning such as artificial neural network, Bayesian neural network, support vector machine in financial distress prediction and comparing their precision of results. Basically, it will assist to widen the use of information mining to predict bank industry distress, but this work gathered information from 18 Bangladesh banks and discovered significant characteristics or financial ratios and tested and evaluated their performance precision on each of these ML methods. The ANN and BNN outperformed SVM better in anticipating distress, according to the consequence.