A bibliometric review of soft computing for recommender systems and sentiment analysis

Soft computing, which focuses on approximate models and provides solutions to complicated real-life issues, has gained increasing momentum in application-specific domains, such as sentiment analysis and recommender systems, to emulate cognitive processes behind decision-making. In this work, bibliometrics and structural topic modeling (STM) were adopted to analyze the text contents of research articles concerning soft computing for sentiment analysis and recommender systems. Results indicated that this research field had experienced a dramatic increase in both quantity and quality as measured by scientific output and their received citations. Using STM, we identified 17 research topics frequently discussed within the analyzed articles. The analysis of annual topic prevalence indicated a shift in research foci from recommender applications to sentiment analysis and a growing interest in soft computing. This study served as a guideline for those seeking to contribute to research on soft computing for sentiment analysis and recommender systems. We also made methodological contributions by combining the leading-edge text mining algorithms to make the time-honored bibliometrics adaptive to the analysis of large quantities of unstructured texts beyond structured publication data statistics.