Research on Semi-supervised Classification with an Ensemble Strategy

DOI

10.2991/icsma-16.2016.119How to use a DOI?

Keywords

sentiment classification; semi-supervised learning; ensemble learning

Abstract

The classification approach based on semi-supervised learning, which is based on small sample sizes marked on the sample by means of a non-labeled improve the classification performance. In order to improve the ability of semi-supervised learning, this paper presents an approach based on the basis of the consistency of the label, the integration of two mainstream semi-supervised classification method is used: Collaborative training methods and label propagation method based on random feature subspace. First, the two semi-supervised learning methods to train the classifier to label unlabeled samples; secondly, select the marked sample rate without labels, so as to obtain a semi-supervised learning method is better than any of the classification results.

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhanhao Han
AU  - Shiqun Yin
PY  - 2016/12
DA  - 2016/12
TI  - Research on Semi-supervised Classification with an Ensemble Strategy
BT  - Proceedings of the 2016 4th International Conference on Sensors, Mechatronics and Automation (ICSMA 2016)
PB  - Atlantis Press
SP  - 681
EP  - 684
SN  - 1951-6851
UR  - https://doi.org/10.2991/icsma-16.2016.119
DO  - 10.2991/icsma-16.2016.119
ID  - Han2016/12
ER  -