Abstract: Being applied to handling large-scale problems, support vector machines(SVMs) needs longer training time and larger memory. I}he paper analyzed the limitation of SVMs, classified the algorithms of applying SVMs to handle large-scale problems into seven types, and made profound and comprehensive analysis of each kind of algorithm. Moreover, some issues valuable for future exploration in this area were indicated and discussed.

Key words: Support vector machines, Largcscale problem, Machine learning