Page 119 - Journal of Library Science in China, Vol.45, 2019
P. 119

118   Journal of Library Science in China, Vol.11, 2019



            evaluate the impacts of monographs.
              With the increasing convenience of full text data acquisition of academic papers, scientometric
            research based on full-texts has attracted interest of many researchers, especially interdisciplinary
            research in the fields of scientometrics, natural language processing, machine learning, and
            information retrieval. Academia has carried out a series of explorations and applications in the
            fields of full-text citation analysis, automatic review and entity topic extraction. Y. Ding et al. (2014)
            statistically analyzed the different citation locations of 866 academic articles in the Journal of the
            American Society for Information Science and Technology. They divided the citation locations
            into six parts: Abstract, Introduction, Literature Review, Methodology, Result, Conclusion and
            Future Work. They calculated citation frequencies with two different ways by considering the
            phenomenon of multiple references. Hu (2014) revealed some phenomena and characteristics
            of citation behaviors with empirical analysis from the three aspects, including specific citation
            location, citation frequency, and citation content in the body of citing literatures. Currently, most
            citation content analysis is based on citation frequency, citation location or citation sentiment.
            Citation content can be used to identify authors’ citation attitudes and mine citing behaviors, so as
            to further improve accuracy and scientificity of book impact evaluation. However, few existing
            book evaluation research considered the influences of citation content information. In addition,
            most of relevant research on citation content takes English literatures as research objects, and there
            is still a lack of research on Chinese literatures, especially Chinese books.
              Therefore, this paper takes the literatures citing Chinese books (abbreviated as citing
            literatures) as the research object and analyzes citing behaviors based on citation content
            (include the sentence with the citation mark and former two sentences before it and the latter
            two sentences after it). Firstly, we selected five disciplines, including computer science, law,
            medicine, literature and sport science, and collected 399 Chinese books from Amazon.cn.
            Then, we collected full-text data of literatures that cited these books manually and extracted
            corresponding citation contents. 2,288 citation contents were obtained in total. Finally, we
            analyzed citation behavior characteristics from four aspects, including citation locations, citation
            intensities, citation lengths and citation sentiments, so as to provide a basis for subsequent book
            evaluation methods.


            1  Related work


            Citation content information includes citation location, citation frequency, length of citation
            content, author’s citation attitude in citing literatures, which can be used for evaluating and
            analyzing books. This paper summarizes relevant research work about citation content information
            from aspects of citation locations, citation intensities, citation lengths and authors’ citation
            sentiments.
   114   115   116   117   118   119   120   121   122   123   124