Open AccessAnalysis of parameters on author attribution of Spanish electronic short texts
Author(s) -
Mario Crespo
Publication year - 2016
Publication title -
research in corpus linguistics
Language(s) - English
Resource type - Journals
ISSN - 2243-4712
DOI - 10.32714/ricl.04.03
Subject(s) - authorship attribution , attribution , affect (linguistics) , task (project management) , linguistics , computer science , feature (linguistics) , process (computing) , applied linguistics , psychology , natural language processing , artificial intelligence , social psychology , communication , engineering , philosophy , systems engineering , operating system
Forensic Linguistics is the analysis of the language which is related to law, either as evidence or as legal discourse. Authorship attribution is the task of identifying the author of a document when the language is used as evidence in a courtroom, so it will be of interest to police investigators and the wider judicial process. Recent advances in Forensic Linguistics are related to the analysis of texts coming from emails, social networks and messages coming from mobile phones. This work continues previous research and explores how different classification algorithms, the size of the text and the type of linguistic feature used in authorship attribution may affect the results in the authorship attribution of Spanish short messages on online forums. Important differences in precision have been assessed when varying both the size of the texts investigated and the algorithms used for classification.