A Comparision of Multiclass Support Vector Machine Algorithms
Zhifeng Hao, Bo Liu, Xiaowei Yang · 2006
Support vector machines (SVMS) were originally designed for binary classifications. For multi-classifications, they are usually converted into binary ones, and up to date, several methods have been proposed to decompose and reconstruct multi-class classification problems. In this paper, we compare the performance of these algorithms. They are applied to eight UCI data sets and the ten-folder method is adopted in our experiments. The results show that the one-against-one scheme is not always better than one-against-all scheme and one-against-all perform just as well as one-against-one approaches