SU@PAN'2015

Experiments in author verification

Stanimir Nikolov, Dobrinka Tabakova, Stefan Savov, Yasen Kiprov, Preslav Nakov

Research output: Contribution to journalArticle

Abstract

We describe the submission of the Sofia University team for the Author Identification Task, part of the PAN 2015 Challenge. Given a small set of documents by a single person and a "questioned" document, possibly of a different genre and/or topic, the task is to determine whether the questioned document was written by the same person who wrote the known document set. This is a hard but realistic formulation of the task, also known as author verification. We experimented with an SVM classifier using variety of features extracted from publicly available resources. Our solution was among the fastest, and running time was an official evaluation metric; however, our results were not so strong on AUC and C1.

Original languageEnglish
JournalUnknown Journal
Volume1391
Publication statusPublished - 2015

Fingerprint

Classifiers
experiment
Experiments
document
resource

Keywords

  • Author identification
  • Forensic linguistics
  • Machine learning
  • Text mining

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Nikolov, S., Tabakova, D., Savov, S., Kiprov, Y., & Nakov, P. (2015). SU@PAN'2015: Experiments in author verification. Unknown Journal, 1391.

SU@PAN'2015 : Experiments in author verification. / Nikolov, Stanimir; Tabakova, Dobrinka; Savov, Stefan; Kiprov, Yasen; Nakov, Preslav.

In: Unknown Journal, Vol. 1391, 2015.

Research output: Contribution to journalArticle

Nikolov, S, Tabakova, D, Savov, S, Kiprov, Y & Nakov, P 2015, 'SU@PAN'2015: Experiments in author verification', Unknown Journal, vol. 1391.
Nikolov S, Tabakova D, Savov S, Kiprov Y, Nakov P. SU@PAN'2015: Experiments in author verification. Unknown Journal. 2015;1391.
Nikolov, Stanimir ; Tabakova, Dobrinka ; Savov, Stefan ; Kiprov, Yasen ; Nakov, Preslav. / SU@PAN'2015 : Experiments in author verification. In: Unknown Journal. 2015 ; Vol. 1391.
@article{5f469f3612f04d16b82fe0de6cbc08f6,
title = "SU@PAN'2015: Experiments in author verification",
abstract = "We describe the submission of the Sofia University team for the Author Identification Task, part of the PAN 2015 Challenge. Given a small set of documents by a single person and a {"}questioned{"} document, possibly of a different genre and/or topic, the task is to determine whether the questioned document was written by the same person who wrote the known document set. This is a hard but realistic formulation of the task, also known as author verification. We experimented with an SVM classifier using variety of features extracted from publicly available resources. Our solution was among the fastest, and running time was an official evaluation metric; however, our results were not so strong on AUC and C1.",
keywords = "Author identification, Forensic linguistics, Machine learning, Text mining",
author = "Stanimir Nikolov and Dobrinka Tabakova and Stefan Savov and Yasen Kiprov and Preslav Nakov",
year = "2015",
language = "English",
volume = "1391",
journal = "JAPCA",
issn = "1073-161X",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - SU@PAN'2015

T2 - Experiments in author verification

AU - Nikolov, Stanimir

AU - Tabakova, Dobrinka

AU - Savov, Stefan

AU - Kiprov, Yasen

AU - Nakov, Preslav

PY - 2015

Y1 - 2015

N2 - We describe the submission of the Sofia University team for the Author Identification Task, part of the PAN 2015 Challenge. Given a small set of documents by a single person and a "questioned" document, possibly of a different genre and/or topic, the task is to determine whether the questioned document was written by the same person who wrote the known document set. This is a hard but realistic formulation of the task, also known as author verification. We experimented with an SVM classifier using variety of features extracted from publicly available resources. Our solution was among the fastest, and running time was an official evaluation metric; however, our results were not so strong on AUC and C1.

AB - We describe the submission of the Sofia University team for the Author Identification Task, part of the PAN 2015 Challenge. Given a small set of documents by a single person and a "questioned" document, possibly of a different genre and/or topic, the task is to determine whether the questioned document was written by the same person who wrote the known document set. This is a hard but realistic formulation of the task, also known as author verification. We experimented with an SVM classifier using variety of features extracted from publicly available resources. Our solution was among the fastest, and running time was an official evaluation metric; however, our results were not so strong on AUC and C1.

KW - Author identification

KW - Forensic linguistics

KW - Machine learning

KW - Text mining

UR - http://www.scopus.com/inward/record.url?scp=84982840713&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84982840713&partnerID=8YFLogxK

M3 - Article

VL - 1391

JO - JAPCA

JF - JAPCA

SN - 1073-161X

ER -