Approximating mutual information for multi-label feature selection

被引：34

作者：

Lee, J. ^{[1
]}

Lim, H. ^{[1
]}

Kim, D. -W. ^{[1
]}

机构：

[1] Chung Ang Univ, Sch Engn & Comp Sci, Seoul 156756, South Korea

来源：

ELECTRONICS LETTERS | 2012年 / 48卷 / 15期

关键词：

Classification (of information);

D O I：

10.1049/el.2012.1600

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

Proposed is a new multi-label feature selection method that captures relationships between features and labels without transforming the problem into single-label classification. Using approximated joint mutual information, the proposed incremental feature selection algorithm provides markedly better classification performance than well-known conventional methods.

引用

页码：929 / 930

页数：2

共 8 条

[1]

[Anonymous], 2010, P ACM SIGKDD

[2]

[Anonymous], 2008, ISMIR

[3] SOME INTERSECTION-THEOREMS FOR ORDERED SETS AND GRAPHS [J].

CHUNG, FRK ;

GRAHAM, RL ;

FRANKL, P ;

SHEARER, JB .

JOURNAL OF COMBINATORIAL THEORY SERIES A, 1986, 43 (01) :23-37

[4]

Doquire G, 2011, LECT NOTES COMPUT SC, V6691, P9, DOI 10.1007/978-3-642-21501-8_2

[5]

Hariharan B., 2010, P 27 INT C MACH LEAR, P1757

[6]

Tsoumakas G, 2011, J MACH LEARN RES, V12, P2411

[7]

Wang H, 2010, LECT NOTES COMPUT SC, V6316, P126, DOI 10.1007/978-3-642-15567-3_10

[8] Feature selection for multi-label naive Bayes classification [J].

Zhang, Min-Ling ;

Pena, Jose M. ;

Robles, Victor .

INFORMATION SCIENCES, 2009, 179 (19) :3218-3229

← 1 →