In this work, we present our approach for the Author Profiling task of PAN 2019. The task is divided into two sub-problems, bot, and gender detection, for two different languages: English and Spanish. For each instance of the problem and each language, we address the problem differently. We use an ensemble architecture to solve the Bot Detection for accounts that write in English and a single SVM for those who write in Spanish. For the Gender detection we use a single SVM architecture for both the languages, but we pre-process the tweets in a different way. Our final models achieve accuracy over the 90% in the bot detection task, while for the gender detection, of 84.17% and 77.61% respectively for the English and Spanish languages.
Bot and gender detection of twitter accounts using distortion and LSA notebook for PAN at CLEF 2019 / Bacciu, A.; La Morgia, M.; Mei, A.; Nemmi, E. N.; Neri, V.; Stefa, J.. - 2380:(2019). (Intervento presentato al convegno 20th Working Notes of CLEF Conference and Labs of the Evaluation Forum, CLEF 2019 tenutosi a Lugano; Switzerland).
Bot and gender detection of twitter accounts using distortion and LSA notebook for PAN at CLEF 2019
Bacciu A.;La Morgia M.
;Mei A.;Nemmi E. N.
;Stefa J.
2019
Abstract
In this work, we present our approach for the Author Profiling task of PAN 2019. The task is divided into two sub-problems, bot, and gender detection, for two different languages: English and Spanish. For each instance of the problem and each language, we address the problem differently. We use an ensemble architecture to solve the Bot Detection for accounts that write in English and a single SVM for those who write in Spanish. For the Gender detection we use a single SVM architecture for both the languages, but we pre-process the tweets in a different way. Our final models achieve accuracy over the 90% in the bot detection task, while for the gender detection, of 84.17% and 77.61% respectively for the English and Spanish languages.File | Dimensione | Formato | |
---|---|---|---|
Bacciu_Bot-and-gender_2019.pdf
accesso aperto
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
Creative commons
Dimensione
240.44 kB
Formato
Adobe PDF
|
240.44 kB | Adobe PDF |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.