论文信息 - Modification of particle swarm algorithm for the problem of the SVM classifier development

Modification of particle swarm algorithm for the problem of the SVM classifier development

The application of particle swarm algorithm and its modifications for the problem of the SVM classifier development is considered. The modification particle swarm algorithm, allowing to reduce the time required for the SVM classifier development during the search of the kernel function type, values of the kernel function parameters and value of the regularization parameter and providing high quality data classification, is offered. The experimental results confirming the efficiency of the modified particle swarm algorithm for the problem of the SVM classifier development are given.

Liliya Demidova | Yulia Sokolova

[1] Choi-Hong Lai,et al. Particle Swarm Optimisation: Classical and Quantum Perspectives , 2011 .

[2] Sayan Mukherjee,et al. Choosing Multiple Parameters for Support Vector Machines , 2002, Machine Learning.

[3] Vladimir Vapnik,et al. Statistical learning theory , 1998 .

[4] L. Demidova,et al. Use of Fuzzy Clustering Algorithms Ensemble for SVM Classifier Development , 2015 .

[5] Kin Keung Lai,et al. Bio-Inspired Credit Risk Analysis , 2008 .