Skip to main navigation Skip to search Skip to main content

Construction and application of PSO-SVM model for personal credit scoring

  • School of Management, Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The parameters of support vector machine (SVM) are crucial to the model's classification performance. Aiming at the randomicity of selecting the parameters in SVM, this paper constructed a PSO-SVM model by using particle swarm optimization (PSO) to search the parameters of SVM. The model was used for personal credit scoring in commercial banks and particles' fitness function was used to control the type II error which costs huger loss to commercial banks. Compared with BP NN, the application results indicate that PSO-SVM gets higher classification accuracy with lower type II error rate and the model shows stronger robustness, which presents more applicable for commercial banks to control personal credit risks.

Original languageEnglish
Title of host publicationComputational Science - ICCS 2007 - 7th International Conference, Proceedings
PublisherSpringer Verlag
Pages158-161
Number of pages4
EditionPART 4
ISBN (Print)9783540725893
DOIs
StatePublished - 2007
Externally publishedYes
Event7th International Conference on Computational Science, ICCS 2007 - Beijing, China
Duration: 27 May 200730 May 2007

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 4
Volume4490 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th International Conference on Computational Science, ICCS 2007
Country/TerritoryChina
CityBeijing
Period27/05/0730/05/07

Keywords

  • Particle swarm optimization
  • Personal credit scoring
  • Support vector machine

Fingerprint

Dive into the research topics of 'Construction and application of PSO-SVM model for personal credit scoring'. Together they form a unique fingerprint.

Cite this