# Multiwavelength cluster mass estimates and machine learning

## TL;DR

This paper explores using machine learning to directly estimate galaxy cluster masses from multiwavelength data, comparing observable contributions and testing simulation-observation consistency.

## Contribution

It introduces a novel approach of combining multiple observables with machine learning for cluster mass estimation, surpassing traditional scaling law methods.

## Key findings

- Machine learning effectively combines multiwavelength data for mass estimates.
- Importance Permutation quantifies each observable's contribution.
- Simulation-observation correlation tests are proposed and explored.

## Abstract

One emerging application of machine learning methods is the inference of galaxy cluster masses. In this note, machine learning is used to directly combine five simulated multiwavelength measurements in order to find cluster masses. This is in contrast to finding mass estimates for each observable, normally by using a scaling relation, and then combining these scaling law based mass estimates using a likelihood. We also illustrate how the contributions of each observable to the accuracy of the resulting mass measurement can be compared via model-agnostic Importance Permutation values. Thirdly, as machine learning relies upon the accuracy of the training set in capturing observables, their correlations, and the observational selection function, and as the machine learning training set originates from simulations, two tests of whether a simulation's correlations are consistent with observations are suggested and explored as well.