The effect of noise and sample size in the performance of an unsupervised feature relevant determination method for manifold learning
Year: 2008
Type: dissertation
Abstract: The research on unsupervised feature selection is scarce in comparison to that for supervised\nmodels, despite the fact that this is an important issue for many clustering\nproblems. An unsupervised feature selection method for general Finite Mixture Models\nwas recently proposed and subsequently extended to Generative Topographic Mapping\n(GTM), a manifold learning constrained mixture model that provides data clustering\nand visualization. Some of the results of previous research on this unsupervised feature\nselection method for GTM suggested that its performance may be affected by insuficient\nsample size and by noisy data. In this thesis, we test in detail such limitations of the\nmethod and outline some techniques that could provide an at least partial solution to\nthe negative effect of the presence of uninformative noise. In particular, we provide a\ndetailed account of a variational Bayesian formulation of feature relevance determination\nfor GTM.
Authors: Velazco Brao, Jorge Sebastián Lozano
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W582572791