TY - GEN
T1 - Automatic discovery of abusive thai language usages in social networks
AU - Tuarob, Suppawong
AU - Mitrpanont, Jarernsri L.
N1 - Publisher Copyright:
© 2017, Springer International Publishing AG.
PY - 2017
Y1 - 2017
N2 - Social networks have become a standard means of communication that allows a massive amount of users to interact and consume information anywhere and anytime. In Thailand, millions of users have access to social networks, a majority of which include young children. The colloquial nature of social media inherently encourages certain expressions of language that do not conform to the standard, some of which may be considered abusive and offensive. Such ill-mannered language fashion has become increasingly used by a large number of Thai social media users. If these abusive languages are exposed to adolescents without proper guidance, they could compulsorily develop a familiar attitude towards such language styles. To address the issue, we present a set of algorithms based on machine learning, that automatically detect abusive Thai language in social networks. Our best results yield 86% f-measure (88.73% precision and 83.53% recall).
AB - Social networks have become a standard means of communication that allows a massive amount of users to interact and consume information anywhere and anytime. In Thailand, millions of users have access to social networks, a majority of which include young children. The colloquial nature of social media inherently encourages certain expressions of language that do not conform to the standard, some of which may be considered abusive and offensive. Such ill-mannered language fashion has become increasingly used by a large number of Thai social media users. If these abusive languages are exposed to adolescents without proper guidance, they could compulsorily develop a familiar attitude towards such language styles. To address the issue, we present a set of algorithms based on machine learning, that automatically detect abusive Thai language in social networks. Our best results yield 86% f-measure (88.73% precision and 83.53% recall).
KW - Abusive language detection
KW - Large scale social networks
KW - Thai natural language processing
UR - https://www.scopus.com/pages/publications/85034109547
U2 - 10.1007/978-3-319-70232-2_23
DO - 10.1007/978-3-319-70232-2_23
M3 - Conference contribution
AN - SCOPUS:85034109547
SN - 9783319702315
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 267
EP - 278
BT - Digital Libraries
A2 - Cunningham, Sally Jo
A2 - Choemprayong, Songphan
A2 - Crestani, Fabio
PB - Springer Verlag
T2 - 19th International Conference on Asia-Pacific Digital Libraries, ICADL 2017
Y2 - 13 November 2017 through 15 November 2017
ER -