Utilizing the H2O AutoML Approach to Predict Hazardous Near-Earth Objects

The universe is replete with various types of objects that need to be studied from time to time, such as stars, asteroids, comets, and a multitude of small astronomical bodies. The small bodies surrounding the Earth are called Near-Earth Objects (NEOs) and may pose a danger to our planet. Therefore, analyzing the attributes and composition of NEOs by utilizing effective machine-learning approaches is considered a crucial mission to detect hazardous near-earth objects and help astrophysics and other scientists figure out the appropriate solutions before the occurrence of this phenomenon. This research focuses on leveraging the H2O AutoML prediction approach, in conjunction with pertinent data features, to accurately identify hazardous NEOs. The H2O AutoML approach has been applied and evaluated using different data-splitting techniques in three different experiments to reach and demonstrate superior performance, which was achieved with an impressive accuracy of 98.27%, precision of 98.37, recall of 98.17%, and F1-score of 98.26%.