{"title":"On Speeding Up Support Vector Machines: Proximity Graphs Versus Random Sampling for Pre-Selection Condensation","authors":"Xiaohua Liu, Juan F. Beltran, Nishant Mohanchandra, Godfried T. Toussaint","country":null,"institution":"","volume":73,"journal":"International Journal of Computer and Information Engineering","pagesStart":133,"pagesEnd":141,"ISSN":"1307-6892","URL":"https:\/\/publications.waset.org\/pdf\/5928","abstract":"Support vector machines (SVMs) are considered to be\r\nthe best machine learning algorithms for minimizing the predictive\r\nprobability of misclassification. However, their drawback is that for\r\nlarge data sets the computation of the optimal decision boundary is a\r\ntime consuming function of the size of the training set. Hence several\r\nmethods have been proposed to speed up the SVM algorithm. Here\r\nthree methods used to speed up the computation of the SVM\r\nclassifiers are compared experimentally using a musical genre\r\nclassification problem. The simplest method pre-selects a random\r\nsample of the data before the application of the SVM algorithm. Two\r\nadditional methods use proximity graphs to pre-select data that are\r\nnear the decision boundary. One uses k-Nearest Neighbor graphs and\r\nthe other Relative Neighborhood Graphs to accomplish the task.","references":null,"publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 73, 2013"}