Abstract:Applications in many domains such as text mining and natural language processing need to deal with high-dimensional data. High-dimensional data may present better clustering characteristics on a selected low-dimensional subspace. Subspace clustering is to project the data onto a low-dimensional subspace before clustering. Traditional subspace clustering methods employ eigenvalue decomposition to find the projection of the input data and perform K-means or kernel K-means to obtain the clustering matrix. This kind of methods is not only inefficient, but also adopts a two-step method to generate an approximate solution. Although Discriminative K-means (DisKmeans) integrates dimensionality reduction and clustering into a joint framework and solves the optimization problem by kernel K-means, such method needs to find the centroids in the kernel space and class labels iteratively and has a square time complexity. Accordingly, in this paper, we propose an algorithm, namely Fast DisKmeans (FDKM), to obtain the cluster indicator matrix in a direct way. Moreover, our proposed method has a linear time complexity, which is a significant reduction compared with the squared time complexity of DisKmeans. We also demonstrate that solving the object function of DisKmeans is equivalent to representing the cluster assignment matrix by a low-dimensional linear mapping of the data. Based on this observation, we propose the second algorithm, namely Iterative Fast DisKmeans (IFDKM), which also has a linear time complexity. A series of experiments were conducted on several datasets, and the experimental results showed the superior performance of FDKM and IFDKM.

Complete discriminant analysis

Computational and Theoretical Analysis of Supervised Dimensionality Reduction

Discriminative analysis of direct linear discriminant analysis for face recognition

A Complete Discriminative Subspace For Robust Face Recognition

A Novel Method for Constructing Low-Dimensional Linear Discriminative Subspace

Data-Dependent Kernel Discriminant Analysis for Feature Extraction and Classification

Constrained Linear Discrimination Analysis and Face Recognition

Quadratic Discriminant Analysis by Projection

A New Reduced-Rank Linear Discriminant Analysis Method and Its Applications

Discriminative analysis algorithm for linear local tangent spaces

Adaptive Quasiconformal Kernel Fisher Discriminant Analysis Via Weighted Maximum Margin Criterion

Subspace Clustering by Directly Solving Discriminative K-means

Optimal Dimensionality Discriminant Analysis and Its Application to Image Recognition

Local Feature Discriminant Projection

Kernel full-space biased discriminant analysis

A NEW ALGORITHM FOR COMPUTING OPTIMAL DISCRIMINANT VECTORS IN CONDITION OF HIGH-DIMENSIONAL DATA

Pivotal Estimation of Linear Discriminant Analysis in High Dimensions

A Generalized Form of Fisher Linear Discriminant Function

FEMDA: a unified framework for discriminant analysis

Relaxed multi-view discriminant analysis

Max-Min Distance Analysis by Using Sequential SDP Relaxation for Dimension Reduction