On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)

Benoît Frénay; Gauthier Doquire; Michel Verleysen

On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)

Benoît Frénay, Gauthier Doquire, Michel Verleysen

Research output: Contribution in Book/Catalog/Report/Conference proceeding › Conference contribution

Abstract

Despite its popularity as a relevance criterion for feature selection, the mutual information can sometimes be inadequate for this task. Indeed, it is commonly accepted that a set of features maximising the mutual information with the target vector leads to a lower probability of misclassification. However, this assumption is in general not true. Justifications and illustrations of this fact are given in this paper.

Original language	English
Title of host publication	Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)
Publisher	i6doc.com
Publication status	Published - 2012
Externally published	Yes

Keywords

ICTEAM:MLAI

Fingerprint

Dive into the research topics of 'On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)'. Together they form a unique fingerprint.

Cite this

Frénay, B., Doquire, G., & Verleysen, M. (2012). On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). In Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012) i6doc.com.

Frénay, Benoît ; Doquire, Gauthier ; Verleysen, Michel. / On the Potential Inadequacy of Mutual Information for Feature Selection : Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). i6doc.com, 2012.

@inproceedings{c45f3bb008274f988fc546bfa4d229cb,

title = "On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)",

abstract = "Despite its popularity as a relevance criterion for feature selection, the mutual information can sometimes be inadequate for this task. Indeed, it is commonly accepted that a set of features maximising the mutual information with the target vector leads to a lower probability of misclassification. However, this assumption is in general not true. Justifications and illustrations of this fact are given in this paper.",

keywords = "ICTEAM:MLAI",

author = "Beno{\^i}t Fr{\'e}nay and Gauthier Doquire and Michel Verleysen",

year = "2012",

language = "English",

booktitle = "Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)",

publisher = "i6doc.com",

}

Frénay, B, Doquire, G & Verleysen, M 2012, On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). in Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). i6doc.com.

On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). / Frénay, Benoît; Doquire, Gauthier; Verleysen, Michel.
Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). i6doc.com, 2012.

Research output: Contribution in Book/Catalog/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - On the Potential Inadequacy of Mutual Information for Feature Selection

T2 - Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)

AU - Frénay, Benoît

AU - Doquire, Gauthier

AU - Verleysen, Michel

PY - 2012

Y1 - 2012

N2 - Despite its popularity as a relevance criterion for feature selection, the mutual information can sometimes be inadequate for this task. Indeed, it is commonly accepted that a set of features maximising the mutual information with the target vector leads to a lower probability of misclassification. However, this assumption is in general not true. Justifications and illustrations of this fact are given in this paper.

AB - Despite its popularity as a relevance criterion for feature selection, the mutual information can sometimes be inadequate for this task. Indeed, it is commonly accepted that a set of features maximising the mutual information with the target vector leads to a lower probability of misclassification. However, this assumption is in general not true. Justifications and illustrations of this fact are given in this paper.

KW - ICTEAM:MLAI

M3 - Conference contribution

BT - Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)

PB - i6doc.com

ER -

Frénay B, Doquire G, Verleysen M. On the Potential Inadequacy of Mutual Information for Feature Selection: Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). In Proceedings of the 20th International Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012). i6doc.com. 2012