Analyzing the relationship between text features and research proposal productivity
arXiv:2005.08254 · doi:10.1007/s11192-021-03926-x
Abstract
Predicting the output of research grants is of considerable relevance to research funding bodies, scientific entities and government agencies. In this study, we investigate whether text features extracted from projects title and abstracts are able to identify productive grants. Our analysis was conducted in three distinct areas, namely Medicine, Dentistry and Veterinary Medicine. Topical and complexity text features were used to identify predictors of productivity. The results indicate that there is a statistically significant relationship between text features and grants productivity, however such a dependence is weak. A feature relevance analysis revealed that the abstract text length and metrics derived from lexical diversity are among the most discriminative features. We also found that the prediction accuracy has a dependence on the considered project language and that topical features are more discriminative than text complexity measurements. Our findings suggest that text features should be used in combination with other features to assist the identification of relevant research ideas.
arXiv commentExperimental results and text features analysis have been updated and improved
References in corpus (16)
- Quantifying Long-Term Scientific Impact
- A systematic comparison of supervised classifiers
- Characterizing and modeling citation dynamics
- Could scientists use Altmetric.com scores to predict longer term citation counts?
- Strong correlations between text quality and complex networks features
- Using network science and text analytics to produce surveys in a scientific topic
- Patterns of authors contribution in scientific manuscripts
- Word sense disambiguation: a complex network approach
- Probing the topological properties of complex networks modeling short written texts
- Classifying informative and imaginative prose using complex networks
- Comparing the writing style of real and artificial papers
- Word sense disambiguation via high order of learning in complex networks
- Concentric network symmetry grasps authors' styles in word adjacency networks
- Reviewers' ratings and bibliometric indicators: hand in hand when assessing over research proposals?
- Three-feature model to reproduce the topology of citation networks and the effects from authors' visibility on their h-index
- Complex networks analysis of language complexity