Page 132 - 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 131

(4) Citation sentiment
According to the statistics of the corresponding citation sentiments about the books, more than
83% citation sentiments about Chinese books were neutral citations. For citations with sentiments,
the authors prefer positive citations. S.B. Liu and K. Ding (2013) analyzed 147,817 citations in the
BMC Bioinformatics. They found that 62.88% of the citations did not show any obvious sentiment
tendencies and were labeled as neutral citations, while positive citations and negative citations
were 33.59% and 3.53% respectively. W. Lu et al. (2014) annotated 673 citation records of 20
articles in fields of “topic model” and found that 96.14% citations were neutral, while positive
citations and negative citations were 1.49% and 2.38% respectively. M.Y. Zhang et al. (2016)
annotated sentiments of 5,320 citation contents in PLoS One, and the results showed that 96.64%
citation contents were neutral, and the proportions of positive citations and negative citations were
2.12% and 1.24% respectively. It suggests that in terms of the sentiments of citation content, books
are similar to non-specific types of references. Namely, most of the citations were neutral citations.
Most researches have proved that in sentiment citations, both books and non-specific types of
references tend to be positively cited.
In summary, although there were slight differences in the feature distributions of citation content
between books and non-specific types of references in citing literatures, the overall distributions
were relatively close. Of course, as the research data in this paper was limited to 399 Chinese
books in the five disciplines, the data sizes, disciplines, and languages of sample data may lead to
differences in conclusions, and the above analysis results may have some limitations. As theories
and methods of this paper did not depend on specific languages, disciplines, or literature types,
they can be also applied to books and academic articles in other languages or other disciplines.
Therefore, the future research needs to be expanded or increased in the field or discipline category,
language category and number of books.

4.2 Differences of the citing behaviors of books in different disciplines

(1) Citation location
There were differences in the distributions of citation locations in different disciplines, which
indicated that the corresponding distributions of citation locations about Chinese books in each
discipline had its own specific tendency. Compared with Methodology and Discussion, the
proportion of each discipline in the Introduction section was higher, and the difference was
smaller. More books were cited in the Methodology section in computer science. Computer science
was the theoretical basis for systematic research on information, and was also the discipline of
implementation and application in computer systems (Y.M.Zhang, X.F. Ma, & Cheng, 2015).
Therefore, literatures in computer science tend to cite methods, models or algorithms. More than
half of the citations in medicine were concentrated in the Discussion section, existing studies have
showed that the Discussion section of medical literatures was an quiet important section (Editorial