Van Heerden, CJBarnard, E2010-12-142010-12-142010-11Van Heerden, CJ and Barnard, E. 2010. Towards understanding the influence of SVM hyperparameters. 21st Annual Symposium of the Pattern Recognition Association of South Africa (PRASA), Stellenbosch, South Africa, 22-23 November 2010, pp 69-74978-0-7992-2470-2http://hdl.handle.net/10204/467521st Annual Symposium of the Pattern Recognition Association of South Africa (PRASA), Stellenbosch, South Africa, 22-23 November 2010In this article, the authors investigate the relationship between SVM hyperparameters for linear and RBF kernels and classification accuracy. The process of finding SVM hyperparameters usually involves a gridsearch, which is both time-consuming and resource-intensive. On large datasets, 10-fold cross-validation grid searches can become intractable without supercomputers or high performance computing clusters. They present theoretical and empirical arguments as to how SVM hyperparameters scale with N, the amount of learning data. By using these arguments, the authors present a simple algorithm for finding approximate hyperparameters on a reduced dataset, followed by a focused line search on the full dataset. Using this algorithm gives comparable results to performing a grid search on complete datasets.enSupport vector machineHyperparametersPRASA 2010Conference PresentationVan Heerden, C., & Barnard, E. (2010). Towards understanding the influence of SVM hyperparameters. PRASA 2010. http://hdl.handle.net/10204/4675Van Heerden, CJ, and E Barnard. "Towards understanding the influence of SVM hyperparameters." (2010): http://hdl.handle.net/10204/4675Van Heerden C, Barnard E, Towards understanding the influence of SVM hyperparameters; PRASA 2010; 2010. http://hdl.handle.net/10204/4675 .TY - Conference Presentation AU - Van Heerden, CJ AU - Barnard, E AB - In this article, the authors investigate the relationship between SVM hyperparameters for linear and RBF kernels and classification accuracy. The process of finding SVM hyperparameters usually involves a gridsearch, which is both time-consuming and resource-intensive. On large datasets, 10-fold cross-validation grid searches can become intractable without supercomputers or high performance computing clusters. They present theoretical and empirical arguments as to how SVM hyperparameters scale with N, the amount of learning data. By using these arguments, the authors present a simple algorithm for finding approximate hyperparameters on a reduced dataset, followed by a focused line search on the full dataset. Using this algorithm gives comparable results to performing a grid search on complete datasets. DA - 2010-11 DB - ResearchSpace DP - CSIR KW - Support vector machine KW - Hyperparameters KW - PRASA 2010 LK - https://researchspace.csir.co.za PY - 2010 SM - 978-0-7992-2470-2 T1 - Towards understanding the influence of SVM hyperparameters TI - Towards understanding the influence of SVM hyperparameters UR - http://hdl.handle.net/10204/4675 ER -