A novel approach to predicting exceptional growth in research

Abstract

The prediction of exceptional or surprising growth in research is an issue with deep roots and few practical solutions. In this study, we develop and validate a novel approach to forecasting growth in highly specific research communities. Each research community is represented by a cluster of papers. Multiple indicators were tested, and a composite indicator was created that predicts which research communities will experience exceptional growth over the next three years. The accuracy of this predictor was tested using hundreds of thousands of community-level forecasts and was found to exceed the performance benchmarks established in Intelligence Advanced Research Projects Activity's (IARPA) Foresight Using Scientific Exposition (FUSE) program in six of nine major fields in science. Furthermore, 10 of 11 disciplines within the Computing Technologies field met the benchmarks. Specific detailed forecast examples are given and evaluated, and a critical evaluation of the forecasting approach is also provided.

AI Research Funding Portfolios and Extreme Growth

“…Using a comprehensive corpus of scholarly literature published worldwide, we aim to infer RCs from these publications and forecast the clusters' growth. In this section, we describe (1) the dataset of scientific publications used in our analysis, (2) the method we use to infer RCs using direct citation links present in the publications, and (3) the method we use to predict growth in each cluster using the cluster forecasting model proposed by Klavans et al ( 2020 ).…”

“…The method for forecasting the publication growth of each RC uses a recently developed forecasting methodology (Klavans et al, 2020 ). This forecasting method is transparent (i.e., the methods are described in the article) and was tested using data from the Scopus database (roughly 50 million papers) and separately using the merged dataset of WOS, DS, MAG, and CNKI.…”

“…Previous work in several areas is relevant to the study reported here. Rather than providing an extensive review, we point to other papers that have substantially reviewed recent work in large-scale clustering of the research literature (Boyack and Klavans, 2019 ), identification of emerging topics (Small et al, 2014 ), forecasting of growth in research topics (Klavans et al, 2020 ), and use of funding acknowledgments (Liu et al, 2020 ).…”

A detailed open access model of the PubMed literature

“…• Current vitality, based on the inverse of mean reference age 29 . This metric is based on the principle that fast growing topics are based on relatively young (rather than old) literature.…”

“…This metric is based on the principle that fast growing topics are based on relatively young (rather than old) literature. • 3 year annualized growth rate from 2016-2019 29 .…”