A data mining application in credit scoring processes of small and medium enterprises commercial corporate customers

Abstract

The constant need to assess loans makes risk evaluation a very important problem for the banking sector. A crucial function of the banks is to fund households and companies from various industries in the economy. Risk is taken by the banks as soon as a loan is given to an entity. Currently, there are sector‐and‐experience based methods of analysis employed by the banks to estimate the risks to be taken. For the credit process, there exist a large number of studies in the literature on scoring individual clients but there are very few studies on scoring small and medium enterprises (SME) commercial corporate customers. In this study, we propose an objective risk measurement method for the lending process of SME commercial corporate customers and performed classification task of data mining by collecting current customer data on credit evaluation process of a bank. For this purpose, we first create a risk measure by looking into the risks identified for existing customers by the analysts of a bank. These scores are used as target variable in the classification process. Then, we extract rules for estimating these scores using Weka software. We used six different algorithms, and compared results in terms of test accuracy, the number of rules, recall, precision and Kappa statistic. We obtained high accuracy rates on real life data by our approach. As a result, we showed that an objective evaluation strategy is possible to use in the lending process for SME commercial corporate customers in the banking system using data mining. This article is categorized under: Application Areas > Business and Industry Technologies > Classification Application Areas > Industry Specific Applications

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 — 23 citations exist upstream.

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