Detection of Pulsars by Classical Machine Learning Algorithms
Deepanshu Beniwal, Abhishek Roy, Himanshu Yadav, Anamika Chauhan · 2021 2nd International Conference for Emerging Technology (INCET) · 2021
This paper compares the classical machine learning algorithms to classify a pulsar (rapidly spinning neutron star) in the HTRU-2 dataset [6]. This dataset was gathered during the survey, called the High Time Resolution Universe (South), and is an imbalanced dataset consisting of more than 17 thousand records. Being highly imbalanced, several methodologies have been implemented (such as sampling, resampling, and feature selection) to overcome the imbalance between target classes (binary in this case) to achieve higher accuracies. The results of the proposed work show that by applying different methodologies, it is possible to overcome the class imbalance of the dataset, which in turn helps the classical machine learning algorithms to obtain promising accuracies.