Extraction of nationality from crime news

Abdulrahman Alkaff, Masnizah Mohd

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Most of the crimes committed today are reported on the Internet through news, blogs and social networking sites. These sources have provided a huge amount of crime data, presenting a need for a means to extract useful information. In this research, the evaluation of Direct and Indirect extraction of nationality from crime news, along with the additional references to identify the nationalities of suspects, victims and witnesses is presented. Named entity recognition using gazetteers and rule-based extraction are used, in addition to co-reference resolution to link references. The proposed approach was evaluated and compared to manual extraction system. The results indicate that the proposed approach is able to extract most information related to nationality from crimes news and identify the additional information references. Its performance proved good, with 55% precision, 96% recall, and 70% f-measure evaluation metrics.

Original languageEnglish
Pages (from-to)304-312
Number of pages9
JournalJournal of Theoretical and Applied Information Technology
Volume54
Issue number2
Publication statusPublished - 2013

Fingerprint

Crime
Social networking (online)
Named Entity Recognition
Social Networking
Blogs
Evaluation
Metric
Internet

Keywords

  • Crime
  • Information extraction
  • Nationality
  • Suspect
  • Victim

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Extraction of nationality from crime news. / Alkaff, Abdulrahman; Mohd, Masnizah.

In: Journal of Theoretical and Applied Information Technology, Vol. 54, No. 2, 2013, p. 304-312.

Research output: Contribution to journalArticle

@article{e6b6fa91099341f4a3a5e9bbe54e549e,
title = "Extraction of nationality from crime news",
abstract = "Most of the crimes committed today are reported on the Internet through news, blogs and social networking sites. These sources have provided a huge amount of crime data, presenting a need for a means to extract useful information. In this research, the evaluation of Direct and Indirect extraction of nationality from crime news, along with the additional references to identify the nationalities of suspects, victims and witnesses is presented. Named entity recognition using gazetteers and rule-based extraction are used, in addition to co-reference resolution to link references. The proposed approach was evaluated and compared to manual extraction system. The results indicate that the proposed approach is able to extract most information related to nationality from crimes news and identify the additional information references. Its performance proved good, with 55{\%} precision, 96{\%} recall, and 70{\%} f-measure evaluation metrics.",
keywords = "Crime, Information extraction, Nationality, Suspect, Victim",
author = "Abdulrahman Alkaff and Masnizah Mohd",
year = "2013",
language = "English",
volume = "54",
pages = "304--312",
journal = "Journal of Theoretical and Applied Information Technology",
issn = "1992-8645",
publisher = "Asian Research Publishing Network (ARPN)",
number = "2",

}

TY - JOUR

T1 - Extraction of nationality from crime news

AU - Alkaff, Abdulrahman

AU - Mohd, Masnizah

PY - 2013

Y1 - 2013

N2 - Most of the crimes committed today are reported on the Internet through news, blogs and social networking sites. These sources have provided a huge amount of crime data, presenting a need for a means to extract useful information. In this research, the evaluation of Direct and Indirect extraction of nationality from crime news, along with the additional references to identify the nationalities of suspects, victims and witnesses is presented. Named entity recognition using gazetteers and rule-based extraction are used, in addition to co-reference resolution to link references. The proposed approach was evaluated and compared to manual extraction system. The results indicate that the proposed approach is able to extract most information related to nationality from crimes news and identify the additional information references. Its performance proved good, with 55% precision, 96% recall, and 70% f-measure evaluation metrics.

AB - Most of the crimes committed today are reported on the Internet through news, blogs and social networking sites. These sources have provided a huge amount of crime data, presenting a need for a means to extract useful information. In this research, the evaluation of Direct and Indirect extraction of nationality from crime news, along with the additional references to identify the nationalities of suspects, victims and witnesses is presented. Named entity recognition using gazetteers and rule-based extraction are used, in addition to co-reference resolution to link references. The proposed approach was evaluated and compared to manual extraction system. The results indicate that the proposed approach is able to extract most information related to nationality from crimes news and identify the additional information references. Its performance proved good, with 55% precision, 96% recall, and 70% f-measure evaluation metrics.

KW - Crime

KW - Information extraction

KW - Nationality

KW - Suspect

KW - Victim

UR - http://www.scopus.com/inward/record.url?scp=84882571108&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84882571108&partnerID=8YFLogxK

M3 - Article

VL - 54

SP - 304

EP - 312

JO - Journal of Theoretical and Applied Information Technology

JF - Journal of Theoretical and Applied Information Technology

SN - 1992-8645

IS - 2

ER -