Recovering Missing Citations in a Scholarly Network: A 2-Step Citation Analysis to Estimate Publication Importance

摘要

Citation relationships between publications are important for assessing the importance of scholarly components (e.g., authors, publications, and venues) within a network. Missing citation metadata in scholarly databases, however, creates problems for classical citation-based ranking algorithms. In the ACM database, for example, 18.5% of publications don't have citation metadata. In this research we propose an innovative, 2-step method of citation analysis, to investigate the importance of publications for which citation data is missing. Preliminary evaluation results show that this method can effectively uncover the importance of publications without using citation metadata.