A tool for literature prioritization based on unsupervised keyphrase detection.
COVID-19 Explorer is a tool for fast and interactive literature prioritization.
We computed keyphrases based on whole texts, offering seamless exploration and ranking of biomedical documents related to COVID-19 domain. The database, underlying this tool is accessible as CORD19.
Keyphrases are computed with RaKUn, an algorithm for graph-based keyphrase detection. Each keyphrase is ranked based on its centrality in the document graph. Multi-term phrases are computed based on ranking paths.
The scores next to each hit (once COVID-19 Explorer is queried) indicate the keyphrase's score. The higher the score, the more relevant this score to a given document. Combinations of keyphrases can also be used for searching. In this situation, the score of the article is computed as sum of all selected keywords found in the article divided by number of selected keywords.
COVID-19 Explorer is a simple and lightweight solution to exploration of COVID-19-related literature. The code for the whole project is freely available at GitLab.
