BRYT: Automated keyword extraction for open datasets

Umair Ahmed, Charalampos Alexopoulos, Marco Piangerelli, Andrea Polini

Abstract: Highlights•Analysis of European data portals revealed a significant lack of representative metadata associated with datasets.•The importance of high-quality metadata, especially keywords, was established to improve data findability on open data portals.•We evaluated different methods for extracting keywords from dataset descriptions to improve findability and proposed a hybrid method, BRYT, which proved to be more efficient.•We also explored which types of descriptions are most effective for extracting representative keywords with particular methodologies.