Clustering in extreme learning machine feature space

• Published in: Neurocomputing (Amsterdam) Vol. 128; pp. 88 - 95
• Main Authors: He, Qing, Jin, Xin, Du, Changying, Zhuang, Fuzhen, Shi, Zhongzhi
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Amsterdam Elsevier B.V 27.03.2014; Elsevier
• Subjects: Algebra; Algorithmics. Computability. Computer arithmetics; Algorithms; Applied sciences; Approximation; Artificial intelligence; Clustering; Computer science; control theory; systems; Connectionism. Neural networks; Data clustering; Data processing. List processing. Character string processing; Elm; ELM feature space; ELM kMeans; ELM NMF clustering; Exact sciences and technology; Extreme learning machine (ELM); Factorization; Learning; Linear and multilinear algebra, matrix theory; Mapping; Mathematical analysis; Mathematics; Memory organisation. Data processing; Neural networks; Nonnegative matrix factorization (NMF); Sciences and techniques of general use; Software; Theoretical computing; Extreme learning machine (ELM); ELM feature space; ELM kMeans; Nonnegative matrix factorization (NMF); ELM NMF clustering; Data clustering; Cluster analysis; Performance evaluation; Extreme learning machine; Cluster; Function kernel; Matrix factorization; Continuous function; Neural network; Matrix decomposition; Modeling; Kernel method; Learning; Kernel function; Non negative matrix; Feedforward neural nets; Classification; Feedforward; Learning algorithm
• ISSN: 0925-2312; 1872-8286
• DOI: 10.1016/j.neucom.2012.12.063

Extreme learning machine (ELM), used for the “generalized” single-hidden-layer feedforward networks (SLFNs), is a unified learning platform that can use a widespread type of feature mappings. In theory, ELM can approximate any target continuous function and classify any disjoint regions; in application, many experiment results have already demonstrated the good performance of ELM. In view of the good properties of the ELM feature mapping, the clustering problem using ELM feature mapping techniques is studied in this paper. Experiments show that the proposed ELM kMeans algorithm and ELM NMF (nonnegative matrix factorization) clustering can get better clustering results than the corresponding Mercer kernel based methods and the traditional algorithms using the original data. Moreover, the proposed methods have the advantage of being more convenient to implementation and computation, as the ELM feature mapping is much simpler than the Mercer kernel function based feature mapping methods.