Disciplinary Boundary for Knowledge Diffusion: A Case Study in Psychology and Computer Science

Taking a study published in different fields as a case, we compare the impact of a study disseminated in different disciplines to reveal the disciplinary boundaries during the knowledge diffusion process. Results show that depending on different diffusion disciplines, the academic impact of the same study can obviously differ at the topic level. From the perspective of citation topics, it can be observed that the research topics triggered by the two articles have been considerably different over years. First, different topics have remained highly popular for a long time in the diffusion of the two articles, such as SVM (topic18) and LBP (topic24) in Article_PSY citations and sparse representation (topic7) and manifold learning (topic11) in Article_CS citations. Second, the point of popular time of topics in different disciplines is different. For example, the popularity time of topics about PCA, SVM, is 2002 in the diffusion of Article_PSY, earlier than that of Article_CS. Third, the focus of research topics affected by the same study diffused in different disciplines is different. For example, the topics of citations of the two articles have different emphases. Lots of topics in citations of Article_PSY are directly related to face recognition methods such as LDA, PCA, and manifold learning, while feature extraction of multimodal data is more abundant in citations of Article_CS such as eye region and image eigenfaces. The in-depth differential analysis of the knowledge flow between disciplines is helpful for us to understand the communication of knowledge at a micro level and more importantly, to reveal the existing disciplinary boundary. In the future, we will further identify research topics in different disciplines and determine how the same knowledge affects disciplines differently