C-Rank: A Link-based Similarity Measure for Scientific Literature Databases

TL;DR

C-Rank is a new link-based similarity measure for scientific papers that considers both in-links and out-links, effectively handling the unique citation patterns in scientific literature databases to improve retrieval accuracy.

Contribution

The paper introduces C-Rank, a novel similarity measure that disregards citation direction and addresses limitations of existing measures in scientific literature databases.

Findings

C-Rank outperforms existing similarity measures in accuracy.

It effectively handles old and recent papers with different citation patterns.

The proposed normalization method improves similarity assessment.

Abstract

As the number of people who use scientific literature databases grows, the demand for literature retrieval services has been steadily increased. One of the most popular retrieval services is to find a set of papers similar to the paper under consideration, which requires a measure that computes similarities between papers. Scientific literature databases exhibit two interesting characteristics that are different from general databases. First, the papers cited by old papers are often not included in the database due to technical and economic reasons. Second, since a paper references the papers published before it, few papers cite recently-published papers. These two characteristics cause all existing similarity measures to fail in at least one of the following cases: (1) measuring the similarity between old, but similar papers, (2) measuring the similarity between recent, but similar papers, and (3) measuring the similarity between two similar papers: one old, the other recent. In this paper, we propose a new link-based similarity measure called C-Rank, which uses both in-link and out-link by disregarding the direction of references. In addition, we discuss the most suitable normalization method for scientific literature databases and propose an evaluation method for measuring the accuracy of similarity measures. We have used a database with real-world papers from DBLP and their reference information crawled from Libra for experiments and compared the performance of C-Rank with those of existing similarity measures. Experimental results show that C-Rank achieves a higher accuracy than existing similarity measures.