Early identification of scientific breakthroughs through outlier analysis based on research entities

Purpose To address the “anomalies” that occur when scientific breakthroughs emerge, this study focuses on identifying early signs and nascent stages of breakthrough innovations from the perspective of outliers, aiming to achieve early identification of scientific breakthroughs in papers. Design/methodology/approach This study utilizes semantic technology to extract research entities from the titles and abstracts of papers to represent each paper’s research content. Outlier detection methods are then employed to measure and analyze the anomalies in breakthrough papers during their early stages. The development and evolution process are traced using literature time tags. Finally, a case study is conducted using the key publications of the 2021 Nobel Prize laureates in Physiology or Medicine. Findings Through manual analysis of all identified outlier papers, the effectiveness of the proposed method for early identifying potential scientific breakthroughs is verified. Research limitations The study’s applicability has only been empirically tested in the biomedical field. More data from various fields are needed to validate the robustness and generalizability of the method. Practical implications This study provides a valuable supplement to current methods for early identification of scientific breakthroughs, effectively supporting technological intelligence decision-making and services. Originality/Value The study introduces a novel approach to early identification of scientific breakthroughs by leveraging outlier analysis of research entities, offering a more sensitive, precise, and fine-grained alternative method compared to traditional citation-based evaluations, which enhances the ability to identify nascent breakthrough innovations.