Aspect Based Sentiment Analysis (ABSA) is a type of a Sentiment Analysis (SA) technique which is used to determine the aspects of the entities and to identify the sentiment of each aspect. Aspect Category detection is one of the sub task in ABSA. For a given set of reviews and a set of predefined aspect categories, the task of Aspect Category detection is to determine the aspect categories which are discussed in each review. Aspect categories are typically coarser than the aspects and these are not necessarily occurring in the sentences as terms. In this paper, it is to investigate the effectiveness of linguistic, Lexicon, vector based and Semantic features for the task of Aspect Category Detection (ACD). Vector based features capture the sentiment and semantic information. In this experiment, the effectiveness of vector based features over text based features is addressed. The deep learning techniques are used to generate vector based features over word based features. The effectiveness of the proposed system is measured using recall, precision and F1-scores. The proposed system has achieved promising results compared with state-of-the-art techniques of ABSA.
Volume 12 | Issue 1
Pages: 398-405
DOI: 10.5373/JARDCS/V12I1/20201920