Exploring, Investigating and Exploiting Sentiment Analysis Systems

With the explosion of social networks, the amount of data produced by online users every day is huge. Sentiment Analysis (SA) serves as a powerful tool capable of extracting opinions and sentiments from textual data. SA tools analyze text to discern the author’s attitude towards a topic, providing insights into human emotions and opinions that guide decision-making across diverse domains. Developed by computer scientists, SA applications encompass product and marketing analysis, monitoring company strategies, financial forecasting, forecasting political movements, combating misinformation, health analysis, supporting education, and analyzing public opinions. Recent studies in SA focus on techniques or applied models. However, further research is needed to develop efficient and precise SA models customized for specific purposes. SA performance is influenced by data quality, feature selection, domain context, and the choice of classification algorithm. We employ the PRISMA methodology to narrow our search scope, focusing on 4,709 papers published by reputable sources such as Springer, Elsevier, IEEE, ACM, and others from 2020 to present. Through iterative review of titles, abstracts, and full texts, we selected 65 papers relevant to our topic, including 59 research papers and 6 survey papers, for in-depth analysis. Our survey provides a systematic, indepth understanding of advancements across domains, applications, datasets, techniques, and future development. It is a valuable resource for identifying knowledge gaps and suggesting potential domains of multidisciplinary application based on data content. Additionally, it offers insights into future research directions.