Title :
Opinion mining for social media and news items in Romanian
Author :
Cardei, Claudia ; Manisor, Filip ; Rebedea, Traian
Author_Institution :
Fac. of Autom. Control & Comput., Univ. Politeh. of Bucharest, Bucharest, Romania
Abstract :
Although opinion mining solutions with good performance exist for English, it is much more difficult to develop an effective solution for other languages. Moreover, even in English the results deteriorate strongly if the texts that are processed are from a variety of contexts (domains, sources, etc.) We present the preliminary results of several experiments with a system that uses machine learning in order to identify opinions in a corpus in Romanian that was manually annotated. This dataset contains texts from a wide range of sources: popular social networking sites, blogs, comments, news providers and online forums.
Keywords :
data mining; learning (artificial intelligence); social networking (online); text analysis; English; Romanian; blogs; comments; machine learning; news items; news providers; online forums; opinion mining; popular social networking sites; social media; Accuracy; Classification algorithms; Data mining; Machine learning algorithms; Media; Support vector machines; Training data;
Conference_Titel :
Systems and Computer Science (ICSCS), 2013 2nd International Conference on
Conference_Location :
Villeneuve d´Ascq
Print_ISBN :
978-1-4799-2020-4
DOI :
10.1109/IcConSCS.2013.6632054
