06-支持向量机方法.ppt

* SVM objective is convex, S3VM objective is non-convex. Finding a solution for semi-supervised SVM is difficult, which has been the focus of S3VM research. Different approaches: SVMlight, rS3VM, continuation S3VM, deterministic annealing, CCCP, Branch and Bound, SDP convex relaxation, etc. The S3VM optimization challenge * Applicable wherever SVMs are applicable, i.e. almost everywhere Clear mathematical framework More modest assumption than generative model or graph-based methods Advantage Disadvantage Optimization difficult Can be trapped in bad local optima Semi-supervised SVM (S3SV