Mining scholarly articles for citation contexts – A pilot study in information science domain

Citations are widely used to measure scholarly impact, reveal knowledge structure, and facilitate information retrieval. However, works are cited for different purposes in different contexts, for example, to provide a context for the current research, to explain concepts/theories, to describe methods/techniques, and to offer a critique. Understanding citation contexts (how citations are used) helps to devise more precise uses of citations. In this pilot study, we adopted a domain-specific approach to analyze citations of the articles from a major venue in information science. A coding scheme is developed to categorize different uses of citations within the domain. Initial results show some promises. Future studies will increase the sample size and further explore automated methods, such as supervised machine learning, for this problem