The Research on an Improved Fast SVM Classification Algorithm

The fast SVM classification algorithm - FCSVM adopt the way to transform of support vector sets, substituting a subset of all the support vector classification for support vector machines, it makes the classification speed improve a lot compared with traditional SVM algorithm under the premise of not losing. In order to obtain the minimum subset of support vector, and to avoid the movement of support vector, the FCSV M algorithm has been improved. It optimizes the support vector in the classification function by using of dichotomy, and provides the necessary and sufficient conditions for the existence of transformation matrix and construction method to reduce the amount of computation. The result of experiments shows that the improved fast classification algorithms considerably reduces the computational complexity and improves the speed of classification, particularly in the circumstances of the large-scale training set and a large number of support vector, its effects become more obvious.