学术论文创新点的识别与检索入口研究——以情报学期刊论文为例

Research on the Identification and Retrieval Entry of Innovation Points of Academic Papers——Taking the Papers of Information Science Journals as an Example

[目的/意义]对学术论文创新点进行识别、组织并用于检索,不仅有利于提高科研效率,促进知识发现与创新,同时也有助于丰富科技论文创新性监测的途径,达到促进科学研究创新的目的。[方法/过程]本研究以情报学期刊论文为例,利用深度学习方法识别学术论文中的创新句,实现学术论文创新内容的多粒度关联。并以多粒度关联成果为数据基础,通过构建创新点检索系统的方式提供创新点检索入口。[结果/结论]本研究探索的方法可以进行更细粒度的论文创新点的识别和组织,为创新情报检索、创新知识发现、学术论文创新性评价提供一定的基础,对科技创新有所贡献。

[Purposeg/Significance]Identifying,organizing,and searching for innovations in academic papers is not only conducive to improving the efficiency of scientific research and promoting knowledge discovery and innovation,but also helping to enrich the ways of innovation monitoring of scientific papers to achieve the purpose of promoting scientific research innovation.[Method/Process]This study takes information science journal papers as an example,uses deep learning methods to identify innovative sentences in academic papers,and realizes the multi-granularity association of the innovative content of academic papers.Based on the multi-granularity associated results,the innovation points retrieval system is constructed to provide the entrance of innovation points retrieval.[Result/Conclusion]The methods explored in this study can identify and organize more fine-grained paper innovation points,provide a certain foundation for innovation information retrieval,innovation knowledge discovery,and academic paper innovation evaluation,which contribute to scientific and technological innovation.