Sentiment Analysis for Reviews in Spanish: Algorithm for Handling the Negation

Samara Gretel Villalba-Osornio, Juan Andrés Pérez-Celis, Luis Villaseñor-Pineda, Manuel Montes-y-Gómez
2015 Research in Computing Science  
Sentiment Analysis (SA) is a challenging area where several problems still are open, one of them is treating the negation. Most of the research focusses on reviews written in English. Therefore, algorithms to deal with the negation in other languages are needed. This work proposes an algorithm to determine the scope of negative particles and to treat the affected words on reviews in Spanish. Our algorithm includes a novel weight scheme which combines the relative frequency of a word with the
more » ... abulary of the class. The experiments show that handling the negation improves the accuracy in Spanish reviews in most cases. Likewise, the proposed weight scheme reaches a maximum accuracy of 67% before treating the negation. Combining our algorithm with the proposed weight scheme yields an accuracy greater than the one reported in the state of the art.
doi:10.13053/rcs-97-1-2 fatcat:hkjz22oh3fhfxjbpmuougk5fke