Predicting Emerging Interdisciplinary Communities with Functional Attributes of Academic Texts
[Objective]This paper explores the diverse characteristics of knowledge network communities to enhance the predicting effectiveness of emerging scientific trends.[Methods]Based on the retrospective growth path of e-Health communities,we proposed a model integrating vocabulary functional attributes to predict emerging trends with diverse features.[Results]In the e-Health field,integrating topic,technical,and other vocabulary functional attribute features can improve the prediction performance of emerging trends.The RF algorithm model,which combines structure,influence,sequence,and attribute features,performed the best.Communities with large vocabulary functional attribute scales,low density,high mediated centrality,and high volatility were more likely to become emerging communities.Sequence features have limited effectiveness in predicting emerging communities,possibly due to the forward-looking impact of emerging communities.[Limitations]The identification results of vocabulary functionality are domain-dependent,and the validity of the conclusions extended to other fields needs further verification.[Conclusions]Fully exploring the fine-grained semantic features of scientific vocabulary can effectively enhance the prediction performance of emerging trends.It provides valuable insights for scientific content evaluation and technology decision support.
NETL
NSTL
万方数据
