A mutual information based feature selection algorithm

Shuang Cang

doi:10.1109/BMEI.2011.6098784

A mutual information based feature selection algorithm

Shuang Cang^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

6 Citations (Scopus)

Abstract

The objective of the eliminating process is to reduce the size of the input feature set and at the same time to retain the class discriminatory information. This paper proposes and evaluates a new feature selection algorithm using information theory which is the mutual information (MI) between combinations of input features and the class instead of mutual information between a single input feature and the class for both continuous-valued and discrete-valued features. Comparison studies of new and previously published classification algorithms indicate that the proposed algorithm is robust, stable and efficient.

Original language	English
Title of host publication	Proceedings - 2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011
Publisher	IEEE
Pages	2241-2245
Number of pages	5
Volume	4
ISBN (Print)	9781424493524
DOIs	https://doi.org/10.1109/BMEI.2011.6098784
Publication status	Published - 12 Dec 2011
Event	2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011 - Shanghai, China Duration: 15 Oct 2011 → 17 Oct 2011

Conference

Conference	2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011
Country/Territory	China
City	Shanghai
Period	15/10/11 → 17/10/11

Keywords

feature ranking
mutual information and classification
optimal feature set

Access to Document

10.1109/BMEI.2011.6098784

Cite this

@inproceedings{1e7f8ec5f441441b821b2fb44deb1d85,

title = "A mutual information based feature selection algorithm",

abstract = "The objective of the eliminating process is to reduce the size of the input feature set and at the same time to retain the class discriminatory information. This paper proposes and evaluates a new feature selection algorithm using information theory which is the mutual information (MI) between combinations of input features and the class instead of mutual information between a single input feature and the class for both continuous-valued and discrete-valued features. Comparison studies of new and previously published classification algorithms indicate that the proposed algorithm is robust, stable and efficient.",

keywords = "feature ranking, mutual information and classification, optimal feature set",

author = "Shuang Cang",

year = "2011",

month = dec,

day = "12",

doi = "10.1109/BMEI.2011.6098784",

language = "English",

isbn = "9781424493524",

volume = "4",

pages = "2241--2245",

booktitle = "Proceedings - 2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011",

publisher = "IEEE",

address = "United States",

note = "2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011 ; Conference date: 15-10-2011 Through 17-10-2011",

}

TY - GEN

T1 - A mutual information based feature selection algorithm

AU - Cang, Shuang

PY - 2011/12/12

Y1 - 2011/12/12

N2 - The objective of the eliminating process is to reduce the size of the input feature set and at the same time to retain the class discriminatory information. This paper proposes and evaluates a new feature selection algorithm using information theory which is the mutual information (MI) between combinations of input features and the class instead of mutual information between a single input feature and the class for both continuous-valued and discrete-valued features. Comparison studies of new and previously published classification algorithms indicate that the proposed algorithm is robust, stable and efficient.

AB - The objective of the eliminating process is to reduce the size of the input feature set and at the same time to retain the class discriminatory information. This paper proposes and evaluates a new feature selection algorithm using information theory which is the mutual information (MI) between combinations of input features and the class instead of mutual information between a single input feature and the class for both continuous-valued and discrete-valued features. Comparison studies of new and previously published classification algorithms indicate that the proposed algorithm is robust, stable and efficient.

KW - feature ranking

KW - mutual information and classification

KW - optimal feature set

UR - https://www.scopus.com/pages/publications/84855817375

U2 - 10.1109/BMEI.2011.6098784

DO - 10.1109/BMEI.2011.6098784

M3 - Conference contribution

AN - SCOPUS:84855817375

SN - 9781424493524

VL - 4

SP - 2241

EP - 2245

BT - Proceedings - 2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011

PB - IEEE

T2 - 2011 4th International Conference on Biomedical Engineering and Informatics, BMEI 2011

Y2 - 15 October 2011 through 17 October 2011

ER -

A mutual information based feature selection algorithm

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this