Fourth Fifth International Conference on Science and Technology Metrics (STMet 2024)
Exploring Citation Dynamics in Computational Linguistics Research: A Citation Context Analysis Approach
Sandra M K, Rupesh Kumar A Department of Library and Information Science, Tumkur University, Tumakuru- 572103
Abstract: Purpose: The present study aimed to explore the contextual and sectional relevance of citations gathered by the scholarly
writings on computational linguistics research using Scite. Further, it investigates the correlation between supporting, contrasting
and mentioning citations with Scopus citations.
Methodology: The study employed conventional and contextual citation analysis tools and techniques to gather the data.
Scientific writings on computational linguistics were exported from Scopus database. Then the contextual citation data was
gathered from smart citation index Scite.ai. Scite is an emerging classified citation database developed using machine
learning, artificial intelligence and natural language processing. A total of 15607 documents with classified citation statements
were included in the present study. The data was further analysed using MS Excel and statistical tests were conducted with the
help of SPSS.
Results: The study found a gradual increase in the number of publications and citations. 15607 scholarly writings together
garnered 3872 supporting, 375 contrasting and 343855 mentioning citations. That is 1.09% agreement, 0.11% disagreement
and 97.23% research engagement is recorded on computational linguistic research. There is a clear domination of mentioning
citations across all the years, indicating the attitude of scholars in neutrally citing a paper without supporting or contrasting the
existing research. Contrasting citations were less compared to supporting and mentioning citations. But still they are relevant
as it represents critical examination and inconstancies associated published research. Among the top ten journals with
highest number of publications, Transactions of the Association for Computational Linguistics exhibits a lead in terms of
supporting and contrasting citations. The Spearman’s Rank Correlation test revealed a strong positive correlation between
Mentioning citations and Scopus citations with Sperman’s rho .839 and p-value .000. Further Scopus citations exhibit a weak
positive correlation with supporting (rho value .336 and p-value .000) and contrasting citations (rho=.153, p-value=.000). The
sectional citation analysis revealed accumulation of large number of citations (6031) in the “Other†section indicating the
diversity followed in scientific communication.
Conclusion: The present investigation illustrated the contextual citation analysis using computational linguistics research.
The study can be extended to author, journal and institutional level to unleash the real research impact. Contextual citation
analysis is a novel approach that redefines the “all that glitters is gold†approach followed in the traditional citation analysis
Keywords: Citation Context Analysis, Computational Linguistics, Scite.ai Exploring Citation Dynamics in Computational Linguistics Research: A Citation Context Analysis Approach
