- Tram Ho
Exoplanets are unpleasant to observe – some are about the size of Neptune, others smaller than Earth. Some exoplanets orbit their host star every few hundred days; others turn around in just 24 hours.
This is not the first time machine learning algorithms have been used to search for planets beyond the Solar System. Astronomical observatories scan the sky every night and collect quite a bit of data. Scientists use algorithms to capture repetitive signals – signals that can prove the existence of a certain planet.
If a celestial body “looks promising”, it becomes a “planetary candidate”. This “candidate” then needs to be clearly identified by different methods. For the first time, scientists use machine learning algorithms to automatically confirm whether an astronomical object might be planets or not.
“To claim the” planetary candidate “, no one has ever used machine learning before. Instead of speaking about which” candidates “have a high probability of becoming a planet, we can point to how much that probability is worth.
Where the probability of “candidate” is not a planet with value less than 1%, we confirm there must be a planet there “- Dr. David Armstrong of the Department of Physics at the University of Warwick concentration.
The scientists said that this approach and current methods can be used to review large data sets. NASA’s TESS Space Telescope alone identified 1835 “planetary candidates”. It is expected that it will find about 13,000 “candidates”.
Source : Genk