Home // IMMM 2016, The Sixth International Conference on Advances in Information Mining and Management // View article

Semi-supervised Learning in the Framework of Data Multiple 1-D Representation

Authors:
Jianzhong Wang

Keywords: Data 1-D representation; regularization; label boosting; ensemble; semi-supervised learning.

Abstract:
The paper develops 1D-based ensemble method for semi-supervised learning (SSL). The method integrates the classifier based on data 1-D representations and label boosting in a serial ensemble. In each stage, the data set is first represented by several 1-D stacks, which preserve the local similarity between data samples. Then, a 1-D ensemble labeler (1DEL) is constructed and used to create a newborn labeled subset from the unlabeled set. United with the subset, the original labeled is boosted for the next learning stage. The boosting process is repeated till the updated labeled set reaches a certain size. Finally, a 1DEL is applied again to build the classifier. The validity and effectiveness of the method are confirmed by experiments. Comparing to several other popular SSL methods, the results of the proposed method are very promising.

Pages: 1 to 4

Copyright: Copyright (c) IARIA, 2016

Publication date: May 22, 2016

Published in: conference

ISSN: 2326-9332

ISBN: 978-1-61208-477-0

Location: Valencia, Spain

Dates: from May 22, 2016 to May 26, 2016