Utilizing Keywords Evolution in Context for Emerging Trend Detection in Scientific Publications
Nhu Khoa Nguyen, Emanuela Boroş, Gaël Lejeune, Antoine Doucet, Thierry Delahaut · 2022
This paper studies the dynamics between how the representation of terms changes through time and its potential emergence as a trending topic in the future. Previous research focused on contrasting directly two of the most recent representations of detected keywords to form a basis for predicting emerging topics. We, thus, propose the Term Context Evolution approach that extends the range of comparison and analyzes multiple continuous previous representations compared to the current one to form a series that reflects the evolution of context around selected terms to evaluate potential emerging topics. We experimented with our approach to abstracts of computer science publications from 1995 to 2012 by first choosing the most frequent terms from the titles and extracting their representations using a Transformer-based pre-trained language model. Our findings reveal that the proposed method outperformed existing models in terms of recall by being able to detect critical emerging trends and points of emergence.