Research Trend Mapping (2019)

TrendNets: Mapping Emerging Research Trends from Dynamic Co-Word Networks via Sparse Representations

Abstract: Visualizing word co-occurrence information extracted from academic texts of journals or proceedings papers is a widely used approach for understanding research trends within a scientific domain. In this study, we develop a novel visualization method that highlights temporal fluctuations in word co-occurrence frequencies. Our proposed technique formulates an optimization problem that decomposes the entries of co-occurrence matrices into a smoothly varying component and a transient, rapidly increasing component. The latter is interpreted as representing research topics that experience short-term bursts of activity.

Authors: Marie Katsurai, Shunsuke Ono

Publication venue: Scientometrics

Code

Code（Matlab）