Pareto models for discriminative multiclass linear dimensionality reduction

PATTERN RECOGNITION

卷:48

Article

Abou-Moustafa, Karim T.¹  De la Torre, Fernando²  Ferrie, Frank P.^3,4 

[1] Univ Alberta, Dept Comp Sci, Edmonton, AB T6G 2E8, Canada

[2] Carnegie Mellon Univ, Inst Robot, Pittsburgh, PA 15213 USA

[3] McGill Univ, Dept Elect & Comp Engn, Montreal, PQ H3A 2E9, Canada

[4] McGill Univ, Ctr Intelligent Machines, Montreal, PQ H3A 2E9, Canada

关键词: Fisher discriminant analysis;    Supervised linear dimensionality reduction;    Feature transformation;    Metric learning;    Subspace learning;    Multiobjective optimization;    Pareto optimality;    Kullback-Leibler divergence;   

DOI  :  10.1016/j.patcog.2014.11.008

来源: Elsevier

【 摘 要 】

We address the class masking problem in multiclass linear discriminant analysis (LDA). In the multiclass setting, LDA does not maximize each pairwise distance between classes, but rather maximizes the sum of all pairwise distances. This results in serious overlaps between classes that are close to each other in the input space, and degrades classification performance. Our research proposes Pareto Discriminant Analysis (PARDA); an approach for multiclass discriminative analysis that builds over multiobjective optimizing models. PARDA decomposes the multiclass problem to a set of objective functions, each representing the distance between every pair of classes. Unlike existing LDA extensions that maximize the sum of all distances, PARDA maximizes each pairwise distance to maximally separate all class means, while minimizing the class overlap in the lower dimensional space. Experimental results on various data sets show consistent and promising performance of PARDA when compared with well-known multiclass LDA extensions. (C) 2014 Elsevier Ltd. All rights reserved.

【 授权许可】

Free