Applying XGBoost Machine Learning Model to Succor Astronomers Detect Exoplanets in Distant Galaxies

The time when TRAPPIST-1 news became official on 22.02.2017, detection of planets beyond Milky Way Galaxy or planets orbiting around their own sun-like stars became one of the burning topics unlike prior times. There are seven famous exoplanets in TRAPPIST-1 system which are just forty light-years distant, and are available to be explored by our planets and other spacial telescopes. But several thousand other exoplanets are known to astronomers whose habitability is misleading as there is no evidence about contrasting effects take place between these bright stars and their suspected exoplanets. Since majority of the exoplanets are found using transit principle method, so in this research paper a new tool using XG Boost supervised Machine Learning Model is proposed to detect their presence. The results show that the prediction accuracy, precision and F1-score of this model is very high as compared to the other methods used in literature till now. This work is novel as till now no research work implements XGBoost based model of Machine Learning with highly accurate predictive power. None of the previous work has taken care of all the steps of data pre-processing and handling imbalanced data.