THE ANALYSIS OF BANKING DATA FOR PREDICTING BAD LOANS USING KNN.

Abstract

Loan default danger or credit danger assessment is critical to economic establishments which give loans to organizations and people. Loans deliver the chance of being defaulted. To recognize the danger levels of credit score customers (groups and individuals), credit score carriers (bankers) normally collect big amounts of facts on debtors. Statistical predictive analytic techniques may be used to analyse or to decide the hazard tiers concerned in loans. This paper aims to address the question of default prediction of quick-time period loans for a Tunisian business bank. Design/technique/approach The authors have used a database of 924 documents of credit granted to business Tunisian businesses by means of a business financial institution within the years2003, 2004, 2005 and 2006. The naive Bayesian classifier algorithm was used, and the results show that the good category price is of the order of sixty three.85 in keeping with cent. The default opportunity is defined with the aid of the variables measuring operating capital, leverage, solvency, profitability and cash glide signs. Findings The effects of the validation take a look at display that the coolest category fee is of the order of fifty eight.66 according to cent; nonetheless, the mistake types I and II remain rather high at 42.Forty two and 40.47 per cent, respectively. A receiver operating feature curve is plotted to evaluate the performance of the version. The end result shows that the vicinity below the curve criterion is of the order of 69 according to cent. Originality/fee The paper highlights the reality that the Tunisian valuable financial institution obliged all commercial banks to conduct a survey observe to collect qualitative data for better credit notation of the borrowers