A class of smooth semi-supervised SVM by difference of convex functions programming and algorithm

摘要

Owing to its wide applicability, semi-supervised learning is an attractive method for using unlabeled data in classification. Applying a new smoothing strategy to a class of continuous semi-supervised support vector machines (S~3VMs), this paper proposes a class of smooth S~3VMs (S~4VMs) without adding new variables and constraints to the corresponding S~3VMs. Moreover, a general framework for solving the S~4VMs is constructed based on robust DC (difference of convex functions) programming. Furthermore, DC optimization algorithms (DCAs) for solving the S~4VMs are investigated. The resulting DCAs converge and only require solving one linear or quadratic program at each iteration. Numerical experiments on some real-world databases demonstrate that the proposed smooth S~3VMs are feasible and effective, and have comparable results as other S~3VMs.