情报科学 ›› 2025, Vol. 43 ›› Issue (9): 99-108.

• 业务研究 • 上一篇    下一篇

跨学科知识元创新组合识别与学术创新机会发现研究

  

  • 出版日期:2025-09-05 发布日期:2025-12-12

  • Online:2025-09-05 Published:2025-12-12

摘要: 【目的/意义】学术创新机会发现是科研人员开展学术研究的前提,从跨学科知识元组合的角度发现学术创 新机会,不仅拓宽了跨学科研究的视角,而且能为研究人员开展规范化的创新研究提供解决方案。【方法/过程】首 先,结合规则匹配和主题提取方法抽取目标学科与源学科学术论文中的问题、方法知识元,建立学科内的“问题-方 法”共现关系。结合共现关系和语义相似度计算,实现跨学科方法知识元的迁移。设计新颖性、融合度、重要性等 跨学科“问题-方法”创新组合测度方法和创新系数指标,识别具有潜在创新机会的“问题-方法”组合。在此基础 上,通过向量合成与凝聚层次聚类,详细分析多样化的跨学科创新机会。【结果/结论】通过实证研究,获得计算机学 科方法知识元与情报学科问题知识元间丰富多样的创新组合关系。【创新/局限】将学术创新机会表示为“问题-方 法”组合的形式,便于研究人员对创新机会的理解和接受;对识别出的“问题-方法”创新组合进行聚类,便于发现多 样化的潜在创新机会,给研究人员开展创新研究提供了较大的自主选择空间。由于对计算机领域方法的理解有 限,对得出的创新机会解释不够充分

Abstract: 【Purpose/significance】The discovery of academic innovation opportunities is a prerequisite for researchers to conduct aca⁃ demic studies. Identifying these opportunities from the perspective of interdisciplinary knowledge element combinations not only broadens the perspective of interdisciplinary research, but also provides solutions for researchers to carry out standardized innovative research.【Method/process】Initially, rule matching and topic extraction methods were employed to extract problem and method knowl⁃ edge elements from academic papers in the target discipline and the source discipline. This process established the "problem-method" co-occurrence relationships within disciplines. By integrating these co-occurrence relationships with semantic similarity calculations, the transfer of method knowledge elements across disciplines was facilitated. Measurement methods and innovation coefficient indica⁃ tors were developed for interdisciplinary "problem-method" combinations, focusing on the assessment of novelty, integration, and sig⁃ nificance to identify potential innovation opportunities. Subsequently, a detailed analysis of diverse interdisciplinary innovation oppor⁃ tunities was conducted using vector synthesis and hierarchical clustering techniques.【Result/conclusion】Through empirical research, this paper presents a rich and varied array of innovation combination relationships between the method knowledge elements of Com⁃ puter Science and the problem knowledge elements of Information Science.【Innovation/limitation】Representing academic innovation opportunities as "problem-method" combinations enhances the clarity and comprehensibility of these opportunities for researchers. Clustering the identified "problem-method" innovative combinations enables the discovery of a range of potential innovation opportu⁃ nities, thereby offering researchers considerable flexibility in their innovative research pursuits. However, limitations in the under⁃ standing of methods within the Computer Science domain may affect the comprehensiveness of the explanations for the identified inno⁃ vation opportunities.