A convex relaxation framework for a class of semi-supervised learning methods and its application in pattern recognition

摘要

Semi-supervised learning has been an attractive research tool for using unlabeled data in pattern recognition. Applying a novel semi-definite programming (SDP) relaxation strategy to a class of continuous semi-supervised support vector machines (S~3VMs), a new convex relaxation framework for the S~3VMs is proposed based on SDP. Compared with other SDP relaxations for S~3VMs, the proposed methods only require solving the primal problems and can implement L_1-norm regularization. Furthermore, the proposed technique is applied directly to recognize the purity of hybrid maize seeds using near-infrared spectral data, from which we find that the proposed method achieves equivalent performance to the exact solution algorithm for solving the S~3VM in different spectral regions. Experiments on several benchmark data sets demonstrate that the proposed convex technique is competitive with other SDP relaxation methods for solving semi-supervised SVMs in generalization.