Anisotropic cosmology using observational datasets: exploring via machine learning approaches

TL;DR

This paper constrains anisotropic cosmological models using observational datasets and explores machine learning techniques like polynomial regression to analyze the Hubble parameter, providing insights into the universe's anisotropic properties.

Contribution

It introduces a combined approach of observational data constraints and machine learning methods to analyze anisotropic cosmological models, highlighting polynomial regression's effectiveness.

Findings

Best-fit parameters for the anisotropic model are estimated using MCMC.

Polynomial regression outperforms other ML techniques in estimating H(z).

Larger datasets improve the understanding of cosmological scenarios.

Abstract

In the current study, we present the observational data constraints on the parameters space for an anisotropic cosmological model of Bianchi I type spacetime in general relativity (GR). For the analysis, we consider observational datasets of Cosmic Chronometers (CC), Baryon Acoustic Oscillation (BAO), and Cosmic Microwave Background Radiation (CMBR) peak parameters. The Markov chain Monte Carlo (MCMC) technique is utilized to constrain the best-fit values of the model parameters. For this purpose, we use the publicly available Python code from CosmoMC and have developed the contour plots with different constraint limits. For the joint dataset of CC, BAO, and CMBR, the parameter's best-fit values for the derived model are estimated as $ H_0 = 69.9\pm 1.4$ km/s/Mpc, $ \Omega_{m0}=0.277^{+0.017}_{-0.015}$, $ \Omega_{\Lambda 0} = 0.722^{+0.015}_{-0.017}$, and $\Omega_{\sigma 0} = 0.0009\pm0.0001$. To estimate $H(z)$, we explore machine learning (ML) techniques like linear regression, Artificial Neural Network (ANN), and polynomial regression and thereafter analyze the results with the theoretically developed $H(z)$ for the proposed model. Among these ML techniques, the polynomial regression exceeds the performance compared to other techniques. Further, we also note that larger dataset provides a better understanding of the cosmological scenario in terms of ML view point.