支持向量机的SMO算法及其自适应改进研究

The Algorithm Research on Support Vector Machine and Adaptive SMO Improvement

提出在SMO算法上应用自适应学习的思想,并利用求解凸二次规划寻优问题的基础上进行改进的研究.研究表明,基于自适应学习的思想对SMO算法进行改进,可使SVM算法更能适应实际应用快速、高效的需求.

Support vector machine （SVM）is an important kind of statistical machine learning algorithm,which SMO algorithm is effective in practical application. SMO is based on support vector machine to solve quadratic progra- mming problem into a set of smaller problems, so as to achieve the minimum serial. The proposed method is applied in SMO algorithm of adaptive learning ideas, and the improvement on the basis of using the optimum solution convex quadratic programming problem. Therefore, SMO algorithm based on the idea of the adaptive learning has been improved SMO. And it will enable the SVM algorithm to adapt to the practical application of fast and efficient needs.