A linguistic approach for determining the topics of Spanish Twitter messages

摘要: The vast number of opinions and reviews provided in Twitter is helpful order to make interesting findings about a given industry, but the huge messages published every day, it important detect relevant ones. In this respect, search functionality not practical tool when we want poll dealing with set general topics. This article presents an approach classify into various We tackle problem from linguistic angle, taking account part-of-speech, syntactic semantic information, showing how language processing techniques should be adapted deal informal present messages. TASS 2013 General corpus, collection tweets that has been specifically annotated perform text analytics tasks, used as dataset our evaluation framework. carry out wide range experiments determine which kinds information have greatest impact on task they combined obtain best-performing system. results lead us conclude relating features by means contextual adds complementary knowledge over pure lexical models, making possible outperform them standard metrics for multilabel classification tasks.