A Fast Classification Algorithm of Support Vector Machines

Proposed in this paper is a fast classification algorithm of support vector machines—FCSVM （fast classification for support vector machines）. After the transformation on the full set of support vectors, a subset of support vectors, which contains fewer support vectors, is used in classification. The speed of classification is much faster than that of conventional SVMs in the condition that the correct rate does not decline. The classification experiments on the UCI database and the face recognition experiments on the FERET face database are done with this algorithm. The experimental results show that it has better performance and partly overcomes the flaw of standard SVMs, which is slow in the process of classification. This algorithm can remarkably reduce the computation and increase the speed of classification, especially in the case of large number of support vectors.