A Decision-Making System Based on Machine Learning for Commercial Credit Limit

This research presents a novel machine learning model that adjusts commercial credit limits based on financial audit data, offering an alternative to traditional models that categorise customers as either 'good' or 'bad.' It identifies key variables for banks' credit rating processes, improving existing methods. The model ensures objectivity by focusing on financial data from independent audits and excluding past behaviour. The study proposes a classification system for credit limits as “increasing” or “decreasing”, aiming to attract new customers. The random forest achieved the highest success rate of 69.40% among the algorithms tested.