Feature Subset Selection and Ranking for Data Dimensionality Reduction

Author

Wei, Hua-Liang ; Billings, Stephen A.

Author_Institution

Dept. of Autom. Control & Syst. Eng., Sheffield Univ.

Volume

29

Issue

1

fYear

2007

Firstpage

162

Lastpage

166

Abstract

A new unsupervised forward orthogonal search (FOS) algorithm is introduced for feature selection and ranking. In the new algorithm, features are selected in a stepwise way, one at a time, by estimating the capability of each specified candidate feature subset to represent the overall features in the measurement space. A squared correlation function is employed as the criterion to measure the dependency between features and this makes the new algorithm easy to implement. The forward orthogonalization strategy, which combines good effectiveness with high efficiency, enables the new algorithm to produce efficient feature subsets with a clear physical interpretation

Keywords

feature extraction; optimisation; search problems; unsupervised learning; data dimensionality reduction; feature selection; feature subset ranking; feature subset selection; high-dimensional data; squared correlation function; unsupervised forward orthogonal search algorithm; Data mining; Extraterrestrial measurements; Feature extraction; Information analysis; Inspection; Libraries; Principal component analysis; Support vector machines; Time measurement; Unsupervised learning; Dimensionality reduction; feature selection; high-dimensional data.; Algorithms; Artificial Intelligence; Cluster Analysis; Data Compression; Image Enhancement; Image Interpretation, Computer-Assisted; Imaging, Three-Dimensional; Pattern Recognition, Automated; Sensitivity and Specificity;

fLanguage

English

Journal_Title

Pattern Analysis and Machine Intelligence, IEEE Transactions on

Publisher

ieee

ISSN

0162-8828

Type

jour

DOI

10.1109/TPAMI.2007.250607

Filename

4016558