A computational analysis of aspect-based sentiment analysis research through bibliometric mapping and topic modeling

With the rising volume of public and consumer engagement on social media platforms, the field of aspect-based sentiment analysis (ABSA) has garnered substantial attention. ABSA contains the systematic extraction of aspects, the analysis of associated sentiments, and the temporal evolution of these sentiments. Researchers have responded to the burgeoning interest by innovating new methodologies and strategies to address specific research challenges, thereby navigating complex scenarios and evolving challenges within ABSA. While existing reviews on ABSA encompass strategies, methods, and applications utilizing survey methodologies, a conspicuous gap exists in literature specifically addressing the development of methodologies and topics and their interaction in ABSA. Furthermore, the application of topic modeling and keyword co-occurrence has been limited in the extant literature. This study conducts a comprehensive overview of the ABSA field by leveraging bibliometrics, topic modeling, social network analysis, and keyword co-occurrence analysis to scrutinize 1325 ABSA research articles spanning the years 2009 to 2023. The analyses encompass research themes and topics, scientific collaborations, top publication sources, research areas, institutions, countries/regions, and publication and citation trends. Beyond examining and contrasting the connections between research topics and methodologies, this study identifies emerging trends and hotspots, providing researchers with insight into technical directions, limitations, and future research regarding ABSA topics and methodologies.