Importancia del lenguaje coloquial y de los símbolos de puntuación en el perfilado de autores

In recent years, author profiling (AP) has become a very relevant task for natural language processing (NLP). The main goal of 43 Research in Computing Science 115 (2016) pp. 43–56; rec. 2016-04-22; acc. 2016-05-15 AP is automatically determine demographic aspects from an author, for example, genre and age. In this paper we present a method for author profiling; particularly, we are interested in determine the rol of colloquial language and the meaning of diverse punctuation marks. Contrary to previous works, our proposal considers each punctuation mark independently and not as a single attribute that covers all marks. Our hypothesis states that the use of certain punctuation marks together with the use of colloquial language can provide relevant information to a automatic classification method. As an aditional contribution, we compiled and made available a dictionary with the colloquial words we use in this paper. The obteined results show that the proposed features allow enhance traditional text representation schemas.