Extrasolar Planet Detection using Machine Learning

Planets outside the solar system were mysterious to astronomers for a long time; astronomers in the past observed these changes in the night sky and noted them. Since then astrophysics has become a data-driven endeavour. The enormous streaming of data collected by scientific projects like the Kepler mission in the process of detecting Extrasolar planets(Any planet that lies beyond our solar system) by telescopes, satellites, and spaceships is becoming too large, which currently relies on human involvement during the selection and classification processes. This makes the analysis and classification of potential Extrasolar Planets a timeconsuming matter. Therefore, there is a need for an automated and unbiased way to identify the Extrasolar Planets. The method used for identifying Extrasolar planets consists of a Convolutional Neural Network model. The model classifies objects denoted as ‘Candidates’ into either ‘confirmed planets’ or ‘false-positives’. Extrasolar Planets Detection proceeds by analyzing the variations in the brightness of a remote star.