Feature selection with rényi min-entropy

Catuscia Palamidessi; Marco Romanelli

doi:10.1007/978-3-319-99978-4_18

Feature selection with rényi min-entropy

Catuscia Palamidessi
, Marco Romanelli

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

Abstract

We consider the problem of feature selection, and we propose a new information-theoretic algorithm for ordering the features according to their relevance for classification. The novelty of our proposal consists in adopting Rényi min-entropy instead of the commonly used Shannon entropy. In particular, we adopt a notion of conditional min-entropy that has been recently proposed in the field of security and privacy, and which is strictly related to the Bayes error. We evaluate our method on two classifiers and three datasets, and we show that it compares favorably with the corresponding one based on Shannon entropy.

Original language	English
Title of host publication	Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings
Editors	Luca Pancioni, Edmondo Trentin, Friedhelm Schwenker
Publisher	Springer Verlag
Pages	226-239
Number of pages	14
ISBN (Print)	9783319999777
DOIs	https://doi.org/10.1007/978-3-319-99978-4_18
Publication status	Published - 1 Jan 2018
Externally published	Yes
Event	8th IAPR TC3 workshop on Artificial Neural Networks for Pattern Recognition, ANNPR 2018 - Siena, Italy Duration: 19 Sept 2018 → 21 Sept 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11081 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	8th IAPR TC3 workshop on Artificial Neural Networks for Pattern Recognition, ANNPR 2018
Country/Territory	Italy
City	Siena
Period	19/09/18 → 21/09/18

Access to Document

10.1007/978-3-319-99978-4_18

Cite this

Palamidessi, C., & Romanelli, M. (2018). Feature selection with rényi min-entropy. In L. Pancioni, E. Trentin, & F. Schwenker (Eds.), Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings (pp. 226-239). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11081 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-99978-4_18

Palamidessi, Catuscia ; Romanelli, Marco. / Feature selection with rényi min-entropy. Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings. editor / Luca Pancioni ; Edmondo Trentin ; Friedhelm Schwenker. Springer Verlag, 2018. pp. 226-239 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{148213366f04463491e2a06db3834e87,

title = "Feature selection with r{\'e}nyi min-entropy",

abstract = "We consider the problem of feature selection, and we propose a new information-theoretic algorithm for ordering the features according to their relevance for classification. The novelty of our proposal consists in adopting R{\'e}nyi min-entropy instead of the commonly used Shannon entropy. In particular, we adopt a notion of conditional min-entropy that has been recently proposed in the field of security and privacy, and which is strictly related to the Bayes error. We evaluate our method on two classifiers and three datasets, and we show that it compares favorably with the corresponding one based on Shannon entropy.",

author = "Catuscia Palamidessi and Marco Romanelli",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2018.; 8th IAPR TC3 workshop on Artificial Neural Networks for Pattern Recognition, ANNPR 2018 ; Conference date: 19-09-2018 Through 21-09-2018",

year = "2018",

month = jan,

day = "1",

doi = "10.1007/978-3-319-99978-4\_18",

language = "English",

isbn = "9783319999777",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "226--239",

editor = "Luca Pancioni and Edmondo Trentin and Friedhelm Schwenker",

booktitle = "Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings",

}

Palamidessi, C & Romanelli, M 2018, Feature selection with rényi min-entropy. in L Pancioni, E Trentin & F Schwenker (eds), Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11081 LNAI, Springer Verlag, pp. 226-239, 8th IAPR TC3 workshop on Artificial Neural Networks for Pattern Recognition, ANNPR 2018, Siena, Italy, 19/09/18. https://doi.org/10.1007/978-3-319-99978-4_18

Feature selection with rényi min-entropy. / Palamidessi, Catuscia; Romanelli, Marco.
Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings. ed. / Luca Pancioni; Edmondo Trentin; Friedhelm Schwenker. Springer Verlag, 2018. p. 226-239 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11081 LNAI).

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

TY - GEN

T1 - Feature selection with rényi min-entropy

AU - Palamidessi, Catuscia

AU - Romanelli, Marco

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2018.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We consider the problem of feature selection, and we propose a new information-theoretic algorithm for ordering the features according to their relevance for classification. The novelty of our proposal consists in adopting Rényi min-entropy instead of the commonly used Shannon entropy. In particular, we adopt a notion of conditional min-entropy that has been recently proposed in the field of security and privacy, and which is strictly related to the Bayes error. We evaluate our method on two classifiers and three datasets, and we show that it compares favorably with the corresponding one based on Shannon entropy.

AB - We consider the problem of feature selection, and we propose a new information-theoretic algorithm for ordering the features according to their relevance for classification. The novelty of our proposal consists in adopting Rényi min-entropy instead of the commonly used Shannon entropy. In particular, we adopt a notion of conditional min-entropy that has been recently proposed in the field of security and privacy, and which is strictly related to the Bayes error. We evaluate our method on two classifiers and three datasets, and we show that it compares favorably with the corresponding one based on Shannon entropy.

U2 - 10.1007/978-3-319-99978-4_18

DO - 10.1007/978-3-319-99978-4_18

M3 - Conference contribution

AN - SCOPUS:85053622535

SN - 9783319999777

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 226

EP - 239

BT - Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings

A2 - Pancioni, Luca

A2 - Trentin, Edmondo

A2 - Schwenker, Friedhelm

PB - Springer Verlag

T2 - 8th IAPR TC3 workshop on Artificial Neural Networks for Pattern Recognition, ANNPR 2018

Y2 - 19 September 2018 through 21 September 2018

ER -

Palamidessi C, Romanelli M. Feature selection with rényi min-entropy. In Pancioni L, Trentin E, Schwenker F, editors, Artificial Neural Networks in Pattern Recognition - 8th IAPR TC3 Workshop, ANNPR 2018, Proceedings. Springer Verlag. 2018. p. 226-239. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-99978-4_18

Feature selection with rényi min-entropy

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this