Implementation of negation handling techniques using modified syntactic rule in Indonesian sentiment analysis

Abstract

Hotels are one of the supporting facilities for tourism in Indonesia. The development of hotels is an essential part of the tourism sector. Hotel development can be performed by knowing public opinion obtained from user reviews. However, direct comment analysis, by reading reviews one by one, becomes difficult due to the vast numbers involved. This problem can be overcome by implementing sentiment analysis. In practice, sentiment analysis has critical issues that can affect performance, i.e., several words have different meanings in different sentence contexts, and the existence of the negation words can invert the meaning of an entire text. Therefore, this research proposes the implementation of negation handling techniques using modified syntactic rules in Naive Bayes based sentiment analysis. The proposed modified syntactic rules of negation handling techniques have been shown to improve sentiment analysis performance by 3.3%; compared to sentiment analysis without negation handling.