DocumentCode :
2207886
Title :
A Binary Decision Diagram-Based One-Class Classifier
Author :
Kutsuna, Takuro
Author_Institution :
Toyota Central R&D Labs. Inc., Nagakute, Japan
fYear :
2010
fDate :
13-17 Dec. 2010
Firstpage :
284
Lastpage :
293
Abstract :
We propose a novel approach for one-class classification problems where a logical formula is used to estimate the region that covers all examples. A formula is viewed as a model that represents a region and is approximated with respect to its hierarchical local densities. The approximation is done quite efficiently via direct manipulations of a binary decision diagram that is a compressed representation of a Boolean formula. The proposed method has only one parameter to be tuned, and the parameter can be selected properly with the help of the minimum description length principle, which requires no labeled training data. In other words, a one-class classifier is generated from an unlabeled training data thoroughly and automatically. Experimental results show that the proposed method works quite well with synthetic data and some realistic data.
Keywords :
approximation theory; binary decision diagrams; pattern classification; unsupervised learning; Boolean formula; approximation method; binary decision diagram; minimum description length principle; one class classifier; parameter tuned; unlabeled training data; binary decision diagram; minimum description length principle; one-class classification;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Data Mining (ICDM), 2010 IEEE 10th International Conference on
Conference_Location :
Sydney, NSW
ISSN :
1550-4786
Print_ISBN :
978-1-4244-9131-5
Electronic_ISBN :
1550-4786
Type :
conf
DOI :
10.1109/ICDM.2010.84
Filename :
5693982
Link To Document :
بازگشت