Predicting the importance of current papers.

Abstract

This article examines how well one can predict the importance of a current paper (a paper that is recently published in the literature). We look at three factors--journal importance, reference importance and author reputation. Citation-based measures of importance are used for all variables. We find that journal importance is the best predictor (explaining 22.3% out of a potential 29.1% of the variance in the data), and that this correlation value varies significantly by discipline. Journal importance is a better predictor of citation in Computer Science than in any other discipline. While the finding supports the present policy of using journal impact statistics as a surrogate for the importance of current papers, it calls into question the present policy of equally weighting current documents in text-based analyses. We suggest that future researchers take into account the expected importance of a document when attempting to describe the cognitive structure of a field.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 33 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.