- author = "A. Tsakonas and G. Dounias",
- title = "Grammar-guided genetic programming for fuzzy rule-based classification in credit management",
- booktitle = "11th meeting of the International Association for Fuzzy-Set Management and Economy",
- year = "2004",
- address = "Reggio Calabria \& Messina, Italy",
- month = "8-9 " # nov,
- keywords = "genetic algorithms, genetic programming",
- isbn13 = "9788882961466",
-
URL = "
https://eprints.bournemouth.ac.uk/17874/",
-
URL = "
https://eprints.bournemouth.ac.uk/17874/1/tsakonas-dounias-sigef-2004.pdf",
-
URL = "http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.149.5637",
- bibsource = "OAI-PMH server at citeseerx.ist.psu.edu",
- contributor = "CiteSeerX",
- language = "en",
- oai = "oai:CiteSeerXPSU:10.1.1.149.5637",
- size = "15 pages",
- abstract = "The study presents a computational intelligent methodology for fuzzy rule-based classification of enterprises into different categories of credit risk. The presented methodology correspond to an approach to the problem of classifying credit applicants, according to the need for reduction of complexity, higher classification accuracy, and comprehensibility of the acquired decision rules. The data used are both of numerical and linguistic nature and they represent a real world problem, that of deciding whether a loan should be granted or not, in respect to financial details of customers applying for that loan, to a private bank of a southern province of the European Union. The techniques involved in the rule-based categorization task are the inductive machine learning and the type-constrained genetic programming. We examine a two-step model, with a sample of 124 enterprises that applied for a loan, each of which is described by 76 (mainly financial) decision variables, and classified to one of the seven predetermined classes. Special attention is given to the comprehensibility and the ease of use for the acquired decision rules. The application of the proposed methods can make the classification task easier and may minimize significantly the amount of required credit data. We consider that the methodology may also give the chance for the extraction of a comprehensible credit management model or even the incorporation of a related decision support system in banking. The overall architecture of the model can be continuously retrained and reformed, by adding every new credit-risk case, becoming more and more accurate and robust classification models over time",
-
notes = "Aristotle University of Thessaloniki, Dept. of
Informatics, Artificial Intelligence and Information
Analysis Lab
broken Nov 2023 http://gandalf.fcee.urv.es/sigef/",
