Research Of Interdisciplinary Knowledge Flow Based On Citation Content

To solve social problems,it is often difficult to conduct comprehensive and in-depth research in a single subject field area,and multidisciplinary efforts are required.Today's scientific research is also increasingly referring to theories and methods from more disciplines,and crossdisciplinary knowledge flows more and more frequently.With the rise of the Open Access Movement,more and more full-text databases are open for free.At the same time,Natural Language Processing and Machine Learning,especially Deep Learning technology continue to develop.These provide a foundation of data and technical for the research of full-text-based interdisciplinary knowledge flow.The citation relationship is an important form that embodies the interdisciplinary,and the content of the citation is the main carrier for the flow of interdisciplinary knowledge.This research selects the citation content sentences in the full text of academic papers as the main research object,extracts the knowledge contained in these sentences,also namely terminology.Then this paper reveals the phenomenon of interdisciplinary and knowledge flow based on the terminology from different disciplines.The methods and results also compared and analyzed with the methods and results based on the matrix of citation numbers of different disciplines.This research quantitatively analyzes the interdisciplinary situation from the perspective of citation content,reveals the phenomenon of interdisciplinary knowledge flow between different disciplines.This paper also provides a new research method for the study of interdisciplinary and knowledge flow based on citation content.This research mainly includes the extraction of terminology from citation contents,interdisciplinary research from different perspectives,and interdisciplinary knowledge flow.In the aspect of extracting terminology from citation contents,this study regards it as the task of extraction of nounc chunks and feature selection.Firstly,we extract citation contents from the full text information,and construct a corpus of citation contents from different disciplines through discipline identification.Secondly,based on the corpus,we use NLP spa Cy for Partof-speech tagging and extraction of nounc chunks.Then we formulate a screening mechanism such as normalization processing and feature selection.And finally,we get a list of terminology in citation contents from various disciplines.In the aspect of interdisciplinary research,firstly,this paper selects five interdisciplinary indexes.Then we conduct comparative analysis and research from two perspectives.Based on the construction of a corpus of citation contents and terminology lists from different disciplines,the first perspective is based on the matrix of cited numbers by different disciplines.The second is on the research perspective proposed in this article,which is to construct a terminology feature matrix based on the terminology lists from different disciplines.We quantify and analyze the interdisciplinary indexes from both two perspectives,and finally compare the results.In the aspect of knowledge flow across disciplines,this article explores the overall knowledge flow phenomenon and the phenomenon in different section positions.Based on the corpus and terminology list constructed in this research,this article first analyzes the content of knowledge flow from the overall perspective.Then we match the terminology to its location information,count the types and distributions in different disciplines at different locations,and perform a fine-grained analysis of the knowledge flow in different locations.Finally,we select specific disciplines for case study of interdisciplinary knowledge flow,and introduce the phenomenon of interdisciplinary knowledge flow in detail.The results show that the terms are widely distributed in the introduction content.The results emphasize the importance of basic disciplines when they are cited,indicating that the citation location has important reference value for evaluating the importance of disciplines.This research aims to construct a corpus of citation contents and terminology list of citation contends form different disciplines to comprehensively analyze the interdisciplinary based on all indexes and interdisciplinary knowledge flow.Besides,we also compare the methods and results from different perspectives.This research provides new ideas and research methods for interdisciplinary related researches.