Pulsar Candidates Classification using Quantum-Classic hibrid generated SVMs

With the growth in the number of pulsar candidates, there is an urgent need to apply automated methods for their processing and classification, such as the use of Artificial Intelligence. The Support Vector Machine (SVM) algorithm has established itself as one of the main Machine Learning algorithms for classification problems. In its original formulation, training an SVM has quadratic complexity, which opens the door to the search for more efficient resolution methods. One way to improve efficiency is through the use of Quantum Computing, where it is possible to obtain algorithms with advantages over classical computers, such as the HHL. This paper presents the classification of pulsar candidates through the use of SVMs generated via a hybrid classical-quantum automated method, where the use of Quantum Machine Learning allowed the achievement of better accuracies than the maximal accuracy obtained from the classical counterpart of the algorithm.