Sentiment Analysis (SA) has achieved significant growth in recent decades among the evolving applications, including social networks and decision-support systems. Recently, the proliferation of data volumes for Dravidian languages has made SA a significant task to retrieve useful information. The prevailing works are focused more on the monolingual analysis and lack a complete study of the recent trends. The research presents a survey on the evolving landscape of SA in Indian languages with a specific emphasis on Dravidian languages. An overview of SA, emphasizing the levels and the purpose of the analysis is presented. Moreover, the generic process is explained with the major advancements of pre-trained architectures applied in recent works. Along with the various classification methodologies, this study analyzes the strengths conforming to the various training databases and languages, mainly focused on sentiment classification. Furthermore, an exhaustive view of the existing challenges and the applications is provided, offering unique linguistic guidance in the field of SA.
Bhat et al. (Thu,) studied this question.