Qiskit Variational Quantum Classifier on the Pulsar Classification Problem

TL;DR

This paper applies a variational quantum classifier to pulsar data, exploring how different data encoding, feature selection, and model parameters affect classification performance in quantum machine learning.

Contribution

It introduces the application of Qiskit-based variational quantum classifiers to pulsar classification, analyzing the impact of various model configurations.

Findings

Performance varies with feature selection methods

Data encoding and ansatz choices influence accuracy

Quantum classifiers show promise for astrophysical data analysis

Abstract

Quantum Machine Learning is a new computational tool that combines the quantum properties from quantum computing with the pattern recognition from machine learning. In this paper, we apply the Variational Quantum Classifier algorithm to the problem of pulsar classification of candidates from the High Time Resolution Universe 2 dataset. We use Qiskit Machine Learning circuits to compare the performance of the model using different feature selection methods, various number of features and training data size. Comparisons on the model from changing the data encoding and ansatz options are also reported. Keywords: Quantum Computing, Quantum Machine Learning, Astrophysics, Pulsars