Isometric Sliced Inverse Regression for Nonlinear Manifolds Learning

doi:10.1007/s11222-012-9330-z

機構典藏 > College of Sciences > Graduate Institute & Department of Mathematics > Journal Article > Item 987654321/78706

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/78706

Title:	Isometric Sliced Inverse Regression for Nonlinear Manifolds Learning
Authors:	Yao, Wei-ting;Wu, Han-ming
Contributors:	淡江大學數學學系
Keywords:	Hierarchical clustering;Isometric feature mapping (ISOMAP);Nonlinear dimension reduction;Nonlinear manifold;Rank-two ellipse seriation;Sliced inverse regression
Date:	2013-09
Issue Date:	2012-10-20 10:41:17 (UTC+8)
Publisher:	New York: Springer New York LLC
Abstract:	Sliced inverse regression (SIR) was developed to find effective linear dimension-reduction directions for exploring the intrinsic structure of the high-dimensional data. In this study, we present isometric SIR for nonlinear dimension reduction, which is a hybrid of the SIR method using the geodesic distance approximation. First, the proposed method computes the isometric distance between data points; the resulting distance matrix is then sliced according to K-means clustering results, and the classical SIR algorithm is applied. We show that the isometric SIR (ISOSIR) can reveal the geometric structure of a nonlinear manifold dataset (e.g., the Swiss roll). We report and discuss this novel method in comparison to several existing dimension-reduction techniques for data visualization and classification problems. The results show that ISOSIR is a promising nonlinear feature extractor for classification applications.
Relation:	Statistics and Computing 23(5), pp.563-576
DOI:	10.1007/s11222-012-9330-z
Appears in Collections:	[Graduate Institute & Department of Mathematics] Journal Article

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	325	View/Open
isoSIR-Yao_Wu.pdf	first page	25Kb	Adobe PDF	636	View/Open

Loading...