Detecting Fake News in URDU using Classical Supervised Machine Learning Methods and Word/Char N-grams

Yaakov HaCohen-Kerner; Natan Manor; Netanel Bashan; Elyasaf Dimant

Detecting Fake News in URDU using Classical Supervised Machine Learning Methods and Word/Char N-grams

Yaakov HaCohen-Kerner, Natan Manor, Netanel Bashan, Elyasaf Dimant

Jerusalem College of Technology

Research output: Contribution to journal › Conference article › peer-review

Abstract

In this paper, we describe our submissions for the UrduFake 2021 track. We tackled the task entitled “Fake News Detection in the Urdu Language". We developed different models using three classical supervised machine learning methods: Support Vector Classifier, Random Forest, and Logistic Regression. Our machine learning models were applied to various sets of character or word n-gram features. Our best submission was an SVC model using 7,500 char trigrams. This model was ranked in 11^th place out of 34 teams that participated in the discussed track.

Original language	English
Pages (from-to)	1162-1167
Number of pages	6
Journal	CEUR Workshop Proceedings
Volume	3159
State	Published - 2021
Externally published	Yes
Event	Working Notes of FIRE - 13th Forum for Information Retrieval Evaluation, FIRE-WN 2021 - Gandhinagar, India Duration: 13 Dec 2021 → 17 Dec 2021

Bibliographical note

Publisher Copyright:
© 2021 Copyright for this paper by the Forum for Information Retrieval Evaluation, December 13-17, 2021, India.

Keywords

Fake news
supervised machine learning
word/char n-grams

Cite this

@article{2fc37bf942c34dbaaa8c883aac941bcc,

title = "Detecting Fake News in URDU using Classical Supervised Machine Learning Methods and Word/Char N-grams",

abstract = "In this paper, we describe our submissions for the UrduFake 2021 track. We tackled the task entitled “Fake News Detection in the Urdu Language{"}. We developed different models using three classical supervised machine learning methods: Support Vector Classifier, Random Forest, and Logistic Regression. Our machine learning models were applied to various sets of character or word n-gram features. Our best submission was an SVC model using 7,500 char trigrams. This model was ranked in 11th place out of 34 teams that participated in the discussed track.",

keywords = "Fake news, supervised machine learning, word/char n-grams",

author = "Yaakov HaCohen-Kerner and Natan Manor and Netanel Bashan and Elyasaf Dimant",

note = "Publisher Copyright: {\textcopyright} 2021 Copyright for this paper by the Forum for Information Retrieval Evaluation, December 13-17, 2021, India.; Working Notes of FIRE - 13th Forum for Information Retrieval Evaluation, FIRE-WN 2021 ; Conference date: 13-12-2021 Through 17-12-2021",

year = "2021",

language = "אנגלית",

volume = "3159",

pages = "1162--1167",

journal = "CEUR Workshop Proceedings",

issn = "1613-0073",

publisher = "CEUR-WS",

}

TY - JOUR

T1 - Detecting Fake News in URDU using Classical Supervised Machine Learning Methods and Word/Char N-grams

AU - HaCohen-Kerner, Yaakov

AU - Manor, Natan

AU - Bashan, Netanel

AU - Dimant, Elyasaf

PY - 2021

Y1 - 2021

N2 - In this paper, we describe our submissions for the UrduFake 2021 track. We tackled the task entitled “Fake News Detection in the Urdu Language". We developed different models using three classical supervised machine learning methods: Support Vector Classifier, Random Forest, and Logistic Regression. Our machine learning models were applied to various sets of character or word n-gram features. Our best submission was an SVC model using 7,500 char trigrams. This model was ranked in 11th place out of 34 teams that participated in the discussed track.

AB - In this paper, we describe our submissions for the UrduFake 2021 track. We tackled the task entitled “Fake News Detection in the Urdu Language". We developed different models using three classical supervised machine learning methods: Support Vector Classifier, Random Forest, and Logistic Regression. Our machine learning models were applied to various sets of character or word n-gram features. Our best submission was an SVC model using 7,500 char trigrams. This model was ranked in 11th place out of 34 teams that participated in the discussed track.

KW - Fake news

KW - supervised machine learning

KW - word/char n-grams

UR - http://www.scopus.com/inward/record.url?scp=85134231904&partnerID=8YFLogxK

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.conferencearticle???

AN - SCOPUS:85134231904

SN - 1613-0073

VL - 3159

SP - 1162

EP - 1167

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

T2 - Working Notes of FIRE - 13th Forum for Information Retrieval Evaluation, FIRE-WN 2021

Y2 - 13 December 2021 through 17 December 2021

ER -

Detecting Fake News in URDU using Classical Supervised Machine Learning Methods and Word/Char N-grams

Abstract

Bibliographical note

Keywords

Other files and links

Fingerprint

Cite this