Page 124 - Journal of Library Science in China, Vol.45, 2019

ZHANG Chengzhi, LI Zhuo, ZHAO Mengyuan, LIU Jiahao & ZHOU Qingqing / Citing behavior of Chinese books based on citation content 123

Table 1. Discipline distribution of books in the citation content corpus
Discipline Number of books Number of citing literature (article) Total number of citations (time)
Computer science 69 262 284
Law 86 408 548
Literature 82 395 614
Medicine 94 480 585
Sport science 68 198 257
Total 399 1,743 2,288

2.3 Data annotation and processing

Based on the manual annotation, we obtained the location information of citation content in full-
text data, and then obtained the intensity, length and sentiment from the citation content corpus.
The annotation scheme is shown in Table 2.

Table 2. Citation content annotation description
Property Category Property description
Introduction The Section that introduces the writing background and purpose of this article
Related work The Section that introduces the related research work of this article
Methodology The Section that introduces the method or means used in this article
Citation location Data The Section that introduces the source of the data in this article
Experiment The Section that describes the experimental process of this article
Discussion The Section that explains and discusses the experimental results of the article
Conclusion The Section that summarizes the research conclusion
The number of times a book is cited/The number of literatures that cited a
Citation intensity S≥1
book
Citation length - The length of the string of the citation content in the citing literature
Positive citation The citation content reflects the positive attitude of the citing author
Citation sentiment Negative citation The citation content reflects the negative attitude of the citing author
Neural citation The citation content does not reflect the emotional attitude of the citing author

(1) Annotation of citation content locations
We divided the citation content locations (abbreviated as citation locations) into seven sections
according to the types of sections: Introduction, Related Work, Data, Methodology, Experiment,
Discussion and Conclusion. In the process of annotation, we found that the writing styles of
different authors of citation literatures in literature and law were quite different, and it was difficult
to judge the citation location information by the section titles. The organizational structures of
articles in other three disciplines were relatively intuitive, and authors’ writing styles were more
similar. Therefore, we only annotated the citation locations of books in three disciplines: computer
science, medicine and sport science, and obtained 1,045 citations with citation location information.