Ranking Papers by their Short-Term Scientific Impact

TL;DR

This paper introduces an attention-based method to rank scientific papers by their short-term impact, measured through near-future citations, outperforming previous approaches across multiple disciplines.

Contribution

We propose a novel attention mechanism modeling researcher behavior to improve short-term impact prediction of scientific papers.

Findings

Our method outperforms previous models in ranking accuracy.

Effective across multiple scientific disciplines.

Captures recent citation trends through attention mechanism.

Abstract

The constantly increasing rate at which scientific papers are published makes it difficult for researchers to identify papers that currently impact the research field of their interest. Hence, approaches to effectively identify papers of high impact have attracted great attention in the past. In this work, we present a method that seeks to rank papers based on their estimated short-term impact, as measured by the number of citations received in the near future. Similar to previous work, our method models a researcher as she explores the paper citation network. The key aspect is that we incorporate an attention-based mechanism, akin to a time-restricted version of preferential attachment, to explicitly capture a researcher's preference to read papers which received a lot of attention recently. A detailed experimental evaluation on four real citation datasets across disciplines, shows that our approach is more effective than previous work in ranking papers based on their short-term impact.