Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/74332
Title: Locality preserving projections with adaptive neighborhood size
Authors: Hu, W 
Cheng, X
Jiang, Y
Choi, KS 
Lou, J
Keywords: Dimensionality reduction
Feature extraction
Locality preserving projections
Neighborhood size
Issue Date: 2017
Publisher: Springer
Source: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics), 2017, v. 10361, p. 223-234 How to cite?
Journal: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics) 
Abstract: Feature extraction methods are widely employed to reduce dimensionality of data and enhance the discriminative information. Among the methods, manifold learning approaches have been developed to detect the underlying manifold structure of the data based on local invariants, which are usually guaranteed by an adjacent graph of the sampled data set. The performance of the manifold learning approaches is however affected by the locality of the data, i.e. what is the neighborhood size for suitably representing the locality? In this paper, we address this issue through proposing a method to adaptively select the neighborhood size. It is applied to the manifold learning approach Locality Preserving Projections (LPP) which is a popular linear reduction algorithm. The effectiveness of the adaptive neighborhood selection method is evaluated by performing classification and clustering experiments on the real-life data sets.
Description: 13th International Conference on Intelligent Computing, ICIC 2017, 7 - 10 August 2017
URI: http://hdl.handle.net/10397/74332
ISBN: 9783319633084
ISSN: 0302-9743
EISSN: 1611-3349
DOI: 10.1007/978-3-319-63309-1_21
Appears in Collections:Conference Paper

Access
View full-text via PolyU eLinks SFX Query
Show full item record

SCOPUSTM   
Citations

1
Last Week
0
Last month
Citations as of May 27, 2020

WEB OF SCIENCETM
Citations

1
Last Week
0
Last month
Citations as of Jun 3, 2020

Page view(s)

95
Citations as of Jun 1, 2020

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.