Similar to the way Google’s search engine ranks pages within the vast world wide web and matches them to your search query, Meta ranks papers, journals, authors and other biomedical concepts, then matches them to your feeds. Both platforms use a mathematical concept called “eigenvector centrality,” which measures the relative importance of the individual parts, or nodes, within a network.
Meta’s predictive algorithm combines many of these entities to determine the most relevant results. For recently published papers, it predicts an article’s impact 3 years from the publication date.
Top papers in your personalized feeds are organized such that recently published papers with the highest rank are displayed toward the top of the feed.