Analyzing the speed of sound in neutron star with machine learning

Sagnik Chatterjee, Harsha Sudhakaran, Ritam Mallick

Published 2023-02-27Version 1

In this work we have analyzed the variation of speed of sound and trace anomaly inside a neutron star using neural network. We construct family of agnostic equation of state by maintaining thermodynamic stability, speed of sound bound with constraints from recent observation of neutron stars. The data of mass and radius serves an input for the neural network and we obtain speed of sound as an output. The speed of sound shows non-monotonic behaviour and the trained data predicts softer equation of state. The neural network predicts stiff equation of state at the centre of intermediate mass stars and massive stars have relatively softer equation of state at their centre. The trace anomaly shows a non-monotonic behaviour for the untrained data however, when we train our data; with neural network the trace anomaly shows monotonic behaviour and the condition of $\Delta \ge 0$ is maintained.