Affiliations: [a] Department of Computer Science, Faculty of Defence Science and Technology, Universiti Pertahanan Nasional Malaysia, Kuala Lumpur, Malaysia | [b] Data Mining and Optimization Research Group, Centre for Artificial Intelligence Technology, Faculty of Information Science and Technology, Universiti Kebangsaan Malaysia, Selangor, Malaysia
Correspondence:
[*]
Corresponding author: Siti Rohaidah Ahmad, Department of Computer Science, Faculty of Defence Science and Technology, Universiti Pertahanan Nasional Malaysia, Kuala Lumpur, Malaysia. E-mail: [email protected].
Abstract: The rapid growth in web development has transformed today’s communication. The combination of features and corresponding sentiment words (SWs) can help produce accurate, meaningful, and high-quality sentiment analysis (SA) results. There are some basic matters in the study of SA that must be understood, namely, the objects or entities that form a key part of the discussion, the characteristics or features of the object, the SWs, and the connection between the features of the object and the SWs. Failure to identify these basic matters can reduce the accuracy and meaning of the SA results. The main objective of this review is to offer an overview of the role and techniques of feature selection (FS), SWs detection, and the identification of the relationship between features and SWs. The main contributions of this review are its sophisticated categorisations of a large number of recent articles related to FS techniques and the detection of SWs. It also highlights the recent trends in the field of SA research. This review will also look at the metaheuristic approach as a FS technique in SA, identify the strengths and weaknesses of existing FS techniques, and analyse the potential of the metaheuristic approach for solving problems that exist in the selection of features in SA.
Keywords: Sentiment analysis, feature selection, sentiment word, ant colony optimization, metaheuristic algorithm
