Fig. 3: Summary of topic model optimization and keywords composition. | Humanities and Social Sciences Communications

Fig. 3: Summary of topic model optimization and keywords composition.

From: Community dynamics and echo chambers: a longitudinal study of the Belt and Road Initiative in the Twittersphere during COVID-19 pandemic

Fig. 3

Right: By using the elbow technique to locate the best trade-off between \(k\) and \({c\_v}\), a model of nine topics and a model of seven topics were found optimal in P1 and P2, respectively. Left: The top four major topics and their ten most representative words were exemplified to analyze the evolution of major public opinions, covering 89% and 96.5% of tweets posted before and during the early stage of the COVID-19 pandemic, respectively (See Figs. S4–S5 and Tables S4–S7 for distribution of topics and keywords).

Back to article page