Efficient tuning of SVM hyperparameters using radius/margin bound and iterative algorithms

IEEE Transactions on Neural Networks - Tập 13 Số 5 - Trang 1225-1229 - 2002
S.S. Keerthi1
1Department of Mechanical Engineering, National University of Singapore, Singapore

Tóm tắt

The paper discusses implementation issues related to the tuning of the hyperparameters of a support vector machine (SVM) with L/sub 2/ soft margin, for which the radius/margin bound is taken as the index to be minimized, and iterative techniques are employed for computing radius and margin. The implementation is shown to be feasible and efficient, even for large problems having more than 10000 support vectors.

Từ khóa

#Support vector machines #Iterative algorithms #Kernel #Support vector machine classification #Quadratic programming #Polynomials #Mechanical engineering #Algorithm design and analysis #Large-scale systems

Tài liệu tham khảo

schölkopf scholkopf, 1995, extracting support data for a given task, 1st Int Conf Knowledge Discovery Data Mining schölkopf scholkopf, 1999, Estimating the support of a high-dimensional distribution 10.1145/355921.355933 vapnik, 1998, Statistical Learning Theory 10.1162/089976600300015042 cristianini, 1999, dynamically adapting kernels in support vector machines, Advances Neural Inform Processing Syst, 10 10.1007/BF00994018 10.1162/089976601300014493 10.1109/72.822516 platt, 1998, fast training of support vector machines using sequential minimal optimization, Advances in Kernel Methods&#x2014 Support Vector Learning platt, 1998, Sequential Minimal Optimization 10.1023/A:1012450327387 rätsch ratsch, 1999, Benchmark Datasets 10.1023/A:1009715923555