Early Discovery of Key Innovative Publications by Analyzing Emerging Topic Trends

Junfeng Wu, Xiangmin Zhou, Guangyan Huang, Borui Cai, Guang-Li Huang, Hui Zheng, Chi-Hung Chi, Jing He

Abstract: The scientific landscape is rapidly evolving with the quick rise of new fields. Researchers must frequently delve into these nascent areas to stay current. Identifying key innovative publications early is crucial, as they lay the groundwork for these fields. Yet, spotting these key innovative publications early is challenging due to irrelevant short-term citations and the initial scarcity of related papers. This paper introduces a novel approach, Analyzing Emerging Topic Trends (AETT), to pinpoint key innovative publications at an early stage by tracking the trajectory of their associated fields. We’ve developed an algorithm to detect emerging topics (i.e., the new fields) prompted by a publication, which helps filter out unrelated citations. Additionally, our trend-based prediction algorithm anticipates a publication’s future impact, overcoming the initial citation lag. Our extensive testing confirms our method’s precision and utility. The Worldwide Emerging Scientific Topic Seer, our open-source tool on GitHub, automates the discovery process for new scientific domains.