New hybrid method for feature selection and classification using meta-heuristic algorithm in credit risk assessment

Autor: Samaneh Mazaheri, Rohollah Moosavi-Tayebi, Mohammad-Reza Feizi-Derakhshi, Fatemeh Alsadat Razeghi, Yashar Zamani-Harghalani, Jalil Nourmohammadi-Khiarak
Rok vydání: 2019
Předmět:
Zdroj: Iran Journal of Computer Science. 3:1-11
ISSN: 2520-8446
2520-8438
DOI: 10.1007/s42044-019-00038-x
Popis: Credit risk is a factor that arises from the failure of the party to the contract. It is one of the most important factors of risk production in banks and financial companies. Still, there is no standard set of features or indices which have been declared through all credit institutions and according to the classification of customers, they are able to do through terms of credit value. In this paper, a meta-heuristic of imperialist competitive algorithm with modified fuzzy min–max classifier (ICA-MFMCN) is offered to identify an optimal subset of features and increased through accuracy classification and scalability through assessment of credit risk. Performance of proposed ICA-MFMCN classification is approved and recognized using a real credit set that has been selected from a UCI dataset. Classification accuracy is comparable for what has been indicated through resources. The experimental which result, in obtaining new classification by utilizing the proposed are promising for future classification are selection processes in assessment of credit risk through retail, indicating that ICA-MFMCN is one of the ways which can be used to add existing data mining techniques.
Databáze: OpenAIRE
Externí odkaz: