The comparative efficiency of algorithms for the construction of support-vector machines for regression reconstruction tasks

Methods for support-vector machine construction do not require additional a priori information and allow for the processing of large datasets; therefore, they are of extreme importance for a range of computational biology tasks. The principal algorithms for support-vector machine construction are reviewed and a comparative analysis of algorithm efficiency is presented. The most efficient algorithms are identified using critical analysis of the results. The algorithms that were recommended are described in sufficient detail to enable practical implementation. © 2015, Pleiades Publishing, Inc.