Predicting Scientific Impact in Astronomy

Project Description: 

This Honours project will explore whether the future scientific impact of new astronomy papers can be predicted using patterns learned from previously published research. Using publicly available metadata from arXiv and citation records from the NASA Astrophysics Data System, the student will construct a dataset of astronomy papers containing abstracts, publication metadata, and citation histories. The project will train predictive models on historical papers to learn relationships between abstract content, early citation signals, and long-term citation impact. The trained model will then be used to estimate the potential impact of newly submitted papers based on their abstracts and available metadata. The outcome will be a prototype impact-prediction model and an analysis of which textual and bibliometric features are most informative for predicting future citations in astronomy literature. No prior experience with large language models is required.(UCT)

Supervisor

University of Cape Town (UCT)

Co-Supervisor

SAAO