Comparison of techniques for measuring research coverage of scientific papers: A case study

摘要

With a plethora of research papers published all throughout the year, the task of measuring the research coverage of a paper has become ever-important. Different methods have been proposed for calculating coverage scores using the references and citations network of papers, based on coccurences techniques and graph ranking algorithms. In this paper, we propose two techniques for measuring coverage based on author-specified keywords in research papers. We evaluate these techniques with HITS-based coverage technique in a simple experiment using four paper-type inclusivity criteria. Results show that the proposed techniques perform better at providing higher scores for diverse, popular and recent papers while the HITS technique was better at identifying survey papers. The proposed techniques will be used in subsequent studies involving a task-based scientific paper recommender system. The theoretical and practical implications of this study are also been discussed in this paper.