Comparing Statistical Measures for Discovering Emerging Terms in Scopus Publications in the Area of Decision Support in Smart City

Discovery of emerging research topics is an important task for scientists, conference organizers, policymakers, and scientific foundations. The paper aims at comparative analysis of statistical models that can be used for discovering emerging terms in a corpus of documents. Three models are evaluated based on calculation of the TF*IDF, TF*PDF and Energy measures. As a case study, a corpus of abstracts of scientific publications related to decision support in smart city is used that was downloaded from Scopus for 20152020. The models are compared and directions of future research to improve the results, namely usage of combinations of models, analysis of synonyms, and usage of additional rules for filtering out non-emerging terms, are identified.