Giant planets have been identified orbiting 290 of the stars within 500 light-years of the sun; the remaining 4,200 stars appeared to be planet-less. However, Natalie Hinkel of the Southwest Research Institute in San Antonio and colleagues think the number with planets could be much higher, based on results from a new Netflix-like algorithm. They applied a machine-learning algorithm, like the one Netflix uses to recommend movies and TV shows, and identified 368 more stars that are 90 percent likely to host giant (so-far undiscovered) exoplanets. From the team’s abstract to the Astrobiology Science Conference: “We train it on the compositions of stars that host planets (i.e., movies you’ve watched before), learn the patterns and overall trends (i.e., that you like a particular genre of movies), then apply those trends to stars with yet-undetected planets (i.e., movies you’ve never seen), and predict the likelihood that those stars host planets (i.e., give you a percentage that you’ll like the movie).” Click here to learn a little more!
Abstract link: https://agu.confex.com/agu/abscicon19/meetingapp.cgi/Paper/481947
Science News article: https://www.sciencenews.org/article/new-algorithm-finds-nearby-stars-could-host-exoplanets