基于引文内容的中文图书被引行为研究  

Citing Behavior of Chinese Books Based on Citation Content

作　　者：章成志[1] 李卓 赵梦圆 柳嘉昊 周清清[1] ZHANG Chengzhi;LI Zhuo;ZHAO Mengyuan;LIU Jiahao;ZHOU Qingqing

机构地区：[1]南京理工大学经济管理学院

出　　处：《中国图书馆学报》2019年第3期96-109,共14页Journal of Library Science in China

基　　金：国家社会科学基金重大项目“情报学学科建设与情报工作未来发展路径研究”(编号:17ZDA291)的研究成果之一~~

摘　　要：从引文内容角度对图书被引行为进行分析,可改善传统依靠被引频次、专家评论等数据进行图书评价的片面性,进一步提高图书评价结果的准确性和科学性。本研究从亚马逊中文网站上选取计算机、法律、医学、文学和体育五个学科领域的中文图书,通过人工采集方式获取图书在施引文献中的引文内容,由此构建包含2288条引文内容的数据集;然后从引用位置、引用强度、引用长度以及引用情感等方面,分析中文图书被引行为,并比较不同学科领域之间的差异。实验结果表明:不同学科领域对中文图书的引用位置具有不同的分布特征,表现出明显的学科差异;引用强度主要在1—3次,文学领域的平均引用强度最高;引用句长度一般在20—160字之间;施引作者对图书的引用情感中,超过80%表现为中性,而含有感情色彩的引用中,正面引用明显多于负面引用。How to accurately assess book impact from massive data collection with uneven qualities is a challenging problem for libraries and researchers. Traditional methods for book impact assessment are based on citation frequency, library holdings or book reviews. Since most citation frequency based methods fail to indicate in depth citation motivation, they cannot support comprehensive assessment of book impact. Hence, this paper used citation contexts to figure out researchers attitudes and behaviors on citations of Chinese books, so as to improve book impact assessment. Specifically, Chinese books from five disciplines were collected: computer science, law, medicine, literature and sport science from Amazon.cn. Then we extracted citation contexts about these Chinese books from each citing literature manually and built a corpus with 2 288 citation contexts. Then, we analyzed citation behaviors over these Chinese books by mining citation locations, citation intensities, citation lengths and citation sentiments. The experimental results showed that: 1) when citing Chinese books, authors from five disciplines had different preference on citation locations. For example, literatures in computer sciences cite books more in Methodology part, while more than half of citations in medicine were in Discussion part. While citation proportions of Introduction were similar among all disciplines. 2)Citation intensities mainly ranged from 1 to 3. More than three-quarter citation intensities were 1 in computer sciences and sport science. Meanwhile, about half of citation intensities in law, medicine and literature were 1. In addition, citations in literature had more high citation intensities. 3) The citation lengths were concentrated between 20 and 160, and showed distribution similarities in some disciplines. Overall, most citation lengths were in the “40-60” interval, accounting for about 23%. 4) Regarding citation sentiments of Chinese books, more than 80% citations were neutral. Compared with negative citations, there were mo

关 键 词：图书被引行为 引文内容 学科差异 图书评价 

分 类 号：G251[文化科学—图书馆学]