Title:	UNSUPERVISED OPTIMAL DISCRIMINANT PLANE WITH UNCORRELATED FEATURES
Author(s):	Su-Qun Cao, Yun-Feng Bu, Xiao-Ming Zuo, Quan-Yin Zhu
ISBN:	978-972-8939-48-9
Editors:	Yingcai Xiao
Year:	2011
Edition:	Single
Keywords:	Optimal discriminant plane; Uncorrelated feature; Feature extraction; Unsupervised pattern
Type:	Short Paper
First Page:	292
Last Page:	296
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	Optimal discriminant plane plays an important role in feature extraction and has been widely used in the pattern recognition field. On this basis, Jin et al. proposed uncorrelated optimal discriminant plane which is based on Fisher criterion function and the conjugated orthogonal constraint of the total-class scatter matrix. This method needs the class information to calculate the Fisher optimal discriminant vector. Thus, it can only be used in supervised pattern. This paper will extend it to the unsupervised case. The proposed method aims to optimize the defined fuzzy Fisher criterion function to figure out an optimal discriminant vector and fuzzy scatter matrices in an unsupervised way. With the conjugated orthogonal constraint of the fuzzy total-class scatter matrix, a novel feature extraction algorithm based on unsupervised optimal discriminant plane with uncorrelated features is proposed. Experimental results on UCI datasets show its validity.
	Go Back