基于多路径分析的跨学科潜在知识组合识别——以引文分析领域为例

Identification of Interdisciplinary Potential Knowledge Combination Based on Multi-path Analysis:Taking the Field of Citation Analysis as an Example

[目的/意义]在提倡“文理交叉”的新文科建设背景下,识别跨学科潜在知识组合,并分析其合作方案的可行性,对于把握跨学科创新方向、推动学科转型与交叉融合具有重要意义。[方法/过程]基于多路径分析和全文知识提取,构建跨学科潜在知识组合合作潜力识别模型,从路径连通性、平衡性、有效潜在路径数三方面反映两知识的合作潜力,并以图书情报学“引文分析”领域为例,全文提取2016—2020年研究文献及其跨学科参考文献关键词进行实证分析。[结果/结论]实证表明,该模型能够通过多组已形成知识路径识别平衡性较好的跨学科潜在知识组合,并发现“引文网络—技术跨越”“相似度分析—激活函数”“聚类分析—扩散曲线”等识别结果具有理论合作可行性和实际应用价值。

[Purpose/significance]Under the background of the new liberal arts construction that advocates the“intersection of arts and sciences”,identifying potential interdisciplinary knowledge combinations and analyzing the feasibility of their coopera-tion proposal is of great significance for grasping the direction of interdisciplinary innovation and promoting disciplinary cross integra-tion transformation.[Method/process]Based on multi-path analysis and full-text knowledge extraction,an interdisciplinary poten-tial knowledge combination identification model is constructed,which reflects the cooperation potential of the two kinds of knowl-edge from the three aspects of path connectivity,balance,and the number of effective potential paths.Taking the“citation analy-sis”field of LIS as an example,the keywords from the 2016-2020 research literature and its interdisciplinary references for empir-ical analysis are extracted.[Result/conclusion]Empirical results show that the model can identify well-balanced interdisciplinary potential knowledge combination through multiple groups of formed knowledge paths.It is found that recognition results such as“ci-tation network&technology leap”,“similarity analysis&activation function”and“cluster analysis&diffusion curve”have theo-retical cooperation feasibility and practical application value.