CREDIT-RISK ASSESSMENT USING SUPPORT VECTORS MACHINE AND MULTILAYER NEURAL NETWORK MODELS: A COMPARATIVE STUDY CASE OF A TUNISIAN BANK

Abstract

Credit risk evaluation or loan default risk evaluation is important to financial institutions which provide loans to businesses and individuals. Credit and loans have risk of being defaulted. To understand risk levels of credit users (corporations and individuals), credit providers (bankers) normally collect vast amount of information on borrowers. Statistical predictive analytic techniques can be used to analyze or to determine risk levels involved in loans. This study contributes to the credit risk evaluation literature in the Middle East and North Africa (MENA) region. We make a comparative analysis of two different statistical methods of classification (artificial neural network and Support Vector Machine). We use a multilayer neural network model and SVM methodology to predict if a particular applicant can be classified as solvent or bankrupt. We use a database of 1435 files of credits granted to industrial Tunisian companies by a Tunisian commercial bank in 2002, 2003, 2004, 2005 and 2006. The results show that the best prediction model is the multilayer neural network model and the best information set is the one combining accrual, cash-flow and collateral variables. The results show that Multilayers Neural Network models outperform the SVM models in terms of global good classification rates and of reduction of Error type I. In fact, the good classification rates are respectively 90.2% (NNM) and 70.13% (SVM) for the in-sample set and the error type I is of the order of 18.55% (NNM) and 29.91% (SVM).

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 15 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.