Translator Disclaimer
28 January 2008 Efficient sequence classification by R2-Kernel
Author Affiliations +
Proceedings Volume 6809, Visualization and Data Analysis 2008; 68090L (2008)
Event: Electronic Imaging, 2008, San Jose, California, United States
A novel kernel, named R2-kernel, is presented for efficient sequence classification based on the Support Vector Machine (SVM). As an intuitive similarity measure, R2 naturally introduces a stop technique in multi-class SVM evaluation, that is, when there exist K support vectors from class i that has similarity beyond a certain threshold with testing sequence X, we can assign sequence X to class i directly and stop the SVM evaluation. The stop technique is seamless integrated into multi-class SVM to improve its evaluation efficiency without negative impact on the performance. Experimental results confirmed the efficiency of the stop technique introduced by the R2-kernel.
© (2008) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Hansheng Lei "Efficient sequence classification by R2-Kernel", Proc. SPIE 6809, Visualization and Data Analysis 2008, 68090L (28 January 2008);

Back to Top