Nonparallel hyperplane support vector machine for PU learning

In this paper, we propose to apply the nonparallel support vector machine (NPSVM) for positive and unlabeled learning problem(PU learning problem) in which only a small positive examples and a large unlabeled examples can be used. Like Biased-SVM, NPSVM treats the unlabeled set as the negative set w...

Full description

Saved in:
Bibliographic Details
Published in2014 10th International Conference on Natural Computation (ICNC) pp. 703 - 708
Main Authors Ying Zhang, XuChan Ju, YingJie Tian
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.08.2014
Subjects
Educational institutions
Kernel
Noise
Power capacitors
Support vector machines
Training
Vectors
Online AccessGet full text

Cover

More Information
Summary:In this paper, we propose to apply the nonparallel support vector machine (NPSVM) for positive and unlabeled learning problem(PU learning problem) in which only a small positive examples and a large unlabeled examples can be used. Like Biased-SVM, NPSVM treats the unlabeled set as the negative set with noise, while NPSVM is modified so that, the first primal problem is constructed such that all the positive points make the contribution to the positive proximal hyperplane; for the second primal problem, only a part of negative points makes the contribution to the negative proximal hyperplane. And we also give the suggestion on the parameters selection. Experimental results show the efficiency of our method for PU learning problem.
ISSN:2157-9555
DOI:10.1109/ICNC.2014.6975922