A new approach for selecting informative features for text classification

Zinnar Ghasem; Ingo Frommholz; Carsten Maple

A new approach for selecting informative features for text classification

Zinnar Ghasem
, Ingo Frommholz
, Carsten Maple

Institute for Research in Applicable Computing (IRAC)

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

1 Citation (Scopus)

Abstract

Selecting useful and informative features to classify text is not only important to decrease the size of the feature space, but as well for the overall performance and precision of machine learning. In this study we propose a new feature selection method called Informative Feature Selector (IFS). Different machine learning algorithms and datasets have been utilised to examine the effectiveness of IFS, and it is compared to well-established methods, namely Information Gain, Odd Ratio, Chi Square, Mutual Information and Class Discriminative Measure. Our experiments show that IFS is able to outperform aforementioned methods and to produce effective and efficient results.

Original language	English
Title of host publication	CEUR Workshop Proceedings
Publisher	CEUR-WS
Pages	382-389
Volume	1458
Edition	Vol 1458
Publication status	Published - 31 Dec 2015
Event	Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015 - Trier, Germany Duration: 7 Oct 2015 → 9 Oct 2015

Conference

Conference	Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015
Country/Territory	Germany
City	Trier
Period	7/10/15 → 9/10/15
Other	Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015 (07/10/2015-09/10/2015, Trier)

Keywords

feature selection
Text classification
Text preprocessing

Access to Document

http://ceur-ws.org/Vol-1458/G07_CRC81_Ghasem.pdf

Handle.net

Persistent link

Cite this

@inproceedings{d6c0f8b6aecd4a4fa12da88078e7f069,

title = "A new approach for selecting informative features for text classification",

abstract = "Selecting useful and informative features to classify text is not only important to decrease the size of the feature space, but as well for the overall performance and precision of machine learning. In this study we propose a new feature selection method called Informative Feature Selector (IFS). Different machine learning algorithms and datasets have been utilised to examine the effectiveness of IFS, and it is compared to well-established methods, namely Information Gain, Odd Ratio, Chi Square, Mutual Information and Class Discriminative Measure. Our experiments show that IFS is able to outperform aforementioned methods and to produce effective and efficient results.",

keywords = "feature selection, Text classification, Text preprocessing",

author = "Zinnar Ghasem and Ingo Frommholz and Carsten Maple",

year = "2015",

month = dec,

day = "31",

language = "English",

volume = "1458",

pages = "382--389",

booktitle = "CEUR Workshop Proceedings",

publisher = "CEUR-WS",

address = "Germany",

edition = "Vol 1458",

note = "Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015 ; Conference date: 07-10-2015 Through 09-10-2015",

}

Ghasem, Z, Frommholz, I & Maple, C 2015, A new approach for selecting informative features for text classification. in CEUR Workshop Proceedings. Vol 1458 edn, vol. 1458, CEUR-WS, pp. 382-389, Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015, Trier, Germany, 7/10/15. <http://ceur-ws.org/Vol-1458/G07_CRC81_Ghasem.pdf>

TY - GEN

T1 - A new approach for selecting informative features for text classification

AU - Ghasem, Zinnar

AU - Frommholz, Ingo

AU - Maple, Carsten

PY - 2015/12/31

Y1 - 2015/12/31

N2 - Selecting useful and informative features to classify text is not only important to decrease the size of the feature space, but as well for the overall performance and precision of machine learning. In this study we propose a new feature selection method called Informative Feature Selector (IFS). Different machine learning algorithms and datasets have been utilised to examine the effectiveness of IFS, and it is compared to well-established methods, namely Information Gain, Odd Ratio, Chi Square, Mutual Information and Class Discriminative Measure. Our experiments show that IFS is able to outperform aforementioned methods and to produce effective and efficient results.

AB - Selecting useful and informative features to classify text is not only important to decrease the size of the feature space, but as well for the overall performance and precision of machine learning. In this study we propose a new feature selection method called Informative Feature Selector (IFS). Different machine learning algorithms and datasets have been utilised to examine the effectiveness of IFS, and it is compared to well-established methods, namely Information Gain, Odd Ratio, Chi Square, Mutual Information and Class Discriminative Measure. Our experiments show that IFS is able to outperform aforementioned methods and to produce effective and efficient results.

KW - feature selection

KW - Text classification

KW - Text preprocessing

M3 - Conference contribution

VL - 1458

SP - 382

EP - 389

BT - CEUR Workshop Proceedings

PB - CEUR-WS

T2 - Learning, Knowledge, Adaptation Workshops, LWA 2015: Knowledge Discovery, Data Mining and Machine Learning, KDML 2015, Knowledge Management, FGWM 2015, Information Retrieval, IR 2015 and Database Systems, FGDB 2015

Y2 - 7 October 2015 through 9 October 2015

ER -

A new approach for selecting informative features for text classification

Abstract

Conference

Keywords

Access to Document

Handle.net

Fingerprint

Cite this