Hsuan-Tien Lin

Home | MOOCs | AIsk | Courses | Research Group | Awards | Publications | Presentations | Programs/Data


Cost-sensitive SVM Implementation with LIBSVM

Please read the README file for copyright information. You are welcomed to discuss with me if you have difficulty applying them for your needs.

Special thanks to Hanhsing Tu and Te-Kang Jan for testing the implementations.

The latest implementation was finished in December 2010 using LIBSVM 2.84. You can download the source code here.

Usage example:

./svm-train -s 5 -l 3.cost_matrix training_file

where 3.cost_matrix can be replaced by 1 (classification), 2 (absolute) or 4.cost_vector; the "5" is for CSOVO and can be replaced by other choices (type svm-train for help). CSOSR and CSOVO are recommended choices for first-hand use.


The program includes several cost-sensitive SVM formulations, including:


Last updated at CST 13:08, October 04, 2023
Please feel free to contact me: htlin.email.png
Valid HTML 4.0!