54
- https://doi.org/10.1137/1.9781611972788.29
Semi-Supervised Classification with Universum
- Apr 24, 2008
- Dan Zhang +3 more
Previous chapter Next chapter Full AccessProceedings Proceedings of the 2008 SIAM International Conference on Data Mining (SDM)Semi-Supervised Classification with UniversumDan Zhang, Jingdong Wang, Fei Wang, and Changshui ZhangDan Zhang, Jingdong Wang, Fei Wang, and Changshui Zhangpp.323 - 333Chapter DOI:https://doi.org/10.1137/1.9781611972788.29PDFBibTexSections ToolsAdd to favoritesExport CitationTrack CitationsEmail SectionsAboutAbstract The Universum data, defined as a collection of “non-examples” that do not belong to any class of interest, have been shown to encode some prior knowledge by representing meaningful concepts in the same domain as the problem at hand. In this paper, we address a novel semi-supervised classification problem, called semi-supervised Universum, that can simultaneously utilize the labeled data, unlabeled data and the Universum data to improve the classification performance. We propose a graph based method to make use of the Universum data to help depict the prior information for possible classifiers. Like conventional graph based semi-supervised methods, the graph regularization is also utilized to favor the consistency between the labels. Furthermore, since the proposed method is a graph based one, it can be easily extended to the multiclass case. The empirical experiments on the USPS and MNIST datasets are presented to show that the proposed method can obtain superior performances over conventional supervised and semi-supervised methods. Previous chapter Next chapter RelatedDetails Published:2008ISBN:978-0-89871-654-2eISBN:978-1-61197-278-8 https://doi.org/10.1137/1.9781611972788Book Series Name:ProceedingsBook Code:PR130Book Pages:1-869
- # Semi-Supervised Classification
- # Semi-supervised Classification Problem
- # Universum Data
- # Previous Chapter
- # MNIST Datasets
- # Semi-supervised Methods
- # Graph Regularization
- # Unlabeled Data
- # Proposed Method
- # Data Mining