Reduced one-against-all method for multiclass SVM classification |
| |
Authors: | M Arun Kumar M Gopal |
| |
Affiliation: | 1. Department of Ecophysiology and Plant Development, Faculty of Biology and Environmental Protection, University of Lodz, 12/16 Banacha Str., 90-237 Lodz, Poland;2. Department of Immunology and Infectious Biology, Faculty of Biology and Environmental Protection, University of Lodz, 12/16 Banacha Str., 90-237 Lodz, Poland;3. Department of Industrial Microbiology and Biotechnology, Faculty of Biology and Environmental Protection, University of Lodz, 12/16 Banacha Str., 90-237 Lodz, Poland;1. Department of Biotechnology Vegetal and Animal Production, Agricultural Science Center, Federal University of São Carlos, Araras, 13600-970, Brazil;2. Itaqui Campus, Federal University of Pampa, Itaqui, 97650-000, Brazil;3. Department of Animal Nutrition and Production, University of São Paulo Pirassununga, 13635-900, Brazil;2. Department of Animal Biosciences, University of Guelph, Guelph, ON N1G 2W1, Canada |
| |
Abstract: | We present an improved version of one-against-all method for multiclass SVM classification based on subset sample selection, named reduced one-against-all, to achieve high performance in large multiclass problems. Reduced one-against-all drastically decreases the computing effort involved in training one-against-all classifiers, without any compromise in classification accuracy. Computational comparisons on publicly available datasets indicate that the proposed method has comparable accuracy to that of conventional one-against-all method, but with an order of magnitude faster. On the largest dataset considered, reduced one-against-all method achieved 50% reduction in computing time over one-against-all method for almost the same classification accuracy. We further investigated reduced one-against-all with linear kernel for multi-label text categorization applications. Computational results demonstrate the effectiveness of the proposed method on both the text corpuses considered. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|