A Machine Learning Approach for the Identification and Estimation of Pulsars upon Statistical Analysis

The life cycle of stars has been an extensive topic of study in the field of astrophysics as it gives us a better understanding of the formation of planets and life. Stars shine by burning their hydrogen fuel in a nuclear fusion reaction. The energy released by this reaction keeps a star from collapsing onto itself due to its own gravity. But, towards the end of its life, a star depletes all of its fuel and there is nothing to help it from collapsing onto its own gravity. A star that weighs about 3 to 4 solar masses forms either a pulsar neutron star or a magnetron neutron star. Study of these stars is very important as it gives better insight into the formation and destruction of stars and planets. This study aims to develop a precise and effective machine learning model for detecting pulsar stars. Efforts have been put into evaluating the performance of a number of well-known machine-learning algorithms including K-Nearest Neighbors, Multiple Adaptive Regression Spline (MARS), neural network classifiers, etc., for identifying Pulsars using R programming language.