# Improved supervised learning methods for EoR parameters reconstruction

## TL;DR

This paper enhances supervised learning techniques, including neural networks and kernel methods, for reconstructing Epoch of Reionization parameters from 21-cm signal data, achieving accuracy comparable to Bayesian inference.

## Contribution

It introduces improved neural network and kernel regression methods for EoR parameter inference, surpassing previous neural network accuracy and approaching Bayesian performance levels.

## Key findings

- Neural networks improved accuracy by an order of magnitude over previous methods.
- Kernel ridge regression further reduces prediction errors by a factor of a few.
- Methods achieve errors within half of the SKA thermal noise confidence level.

## Abstract

Within the next few years, the Square Kilometer Array (SKA) or one of its pathfinders will hopefully provide a detection of the 21-cm signal fluctuations from the Epoch of Reionization (EoR). Then, the main goal will be to accurately constrain the underlying astrophysical parameters. Currently, this is mainly done with Bayesian inference using Markov Chain Monte Carlo sampling. Recently, studies using neural networks trained to performed inverse modelling have shown interesting results. We build on these by improving the accuracy of the predictions using neural network and exploring other supervised learning methods: the kernel and ridge regressions. Based on a large training set of 21-cm power spectra, we compare the performances of these supervised learning methods. When using an un-noised signal as input, we improve on previous neural network accuracy by one order of magnitude and, using local ridge kernel regression, we gain another factor of a few. We then reach a rms prediction error of a few percents of the 1-sigma confidence level due to SKA thermal noise (as estimated with Bayesian inference). This last performance level requires optimizing the hyper-parameters of the method: how to do that perfectly in the case of an unknown signal remains an open question. For an input signal altered by a SKA-type thermal noise, our neural network recovers the astrophysical parameter values with an error within half of the 1$\sigma$ confidence level due to the SKA thermal noise. This accuracy improves to 10$\%$ of the 1$\sigma$ level when using the local ridge kernel regression (with optimized hyper-parameters). We are thus reaching a performance level where supervised learning methods are a viable alternative to determine the best-fit parameters values.