The Overlooked Potential of Generalized Linear Models in Astronomy - I: Binomial Regression

TL;DR

This paper highlights the underused potential of generalized linear models, particularly binomial regression, in astronomy for analyzing binary data, demonstrating their effectiveness and interpretability compared to neural networks.

Contribution

It introduces the application of GLMs to astronomical binary data, showcasing their predictive power and interpretability from both maximum likelihood and Bayesian perspectives.

Findings

GLMs effectively model star formation conditions in cosmological simulations.

GLMs outperform neural networks in binary classification tasks.

Receiver operating characteristic curves validate GLMs' predictive performance.

Abstract

Revealing hidden patterns in astronomical data is often the path to fundamental scientific breakthroughs; meanwhile the complexity of scientific inquiry increases as more subtle relationships are sought. Contemporary data analysis problems often elude the capabilities of classical statistical techniques, suggesting the use of cutting edge statistical methods. In this light, astronomers have overlooked a whole family of statistical techniques for exploratory data analysis and robust regression, the so-called Generalized Linear Models (GLMs). In this paper -- the first in a series aimed at illustrating the power of these methods in astronomical applications -- we elucidate the potential of a particular class of GLMs for handling binary/binomial data, the so-called logit and probit regression techniques, from both a maximum likelihood and a Bayesian perspective. As a case in point, we present the use of these GLMs to explore the conditions of star formation activity and metal enrichment in primordial minihaloes from cosmological hydro-simulations including detailed chemistry, gas physics, and stellar feedback. We predict that for a dark mini-halo with metallicity $\approx 1.3 \times 10^{-4} Z_{\bigodot}$, an increase of $1.2 \times 10^{-2}$ in the gas molecular fraction, increases the probability of star formation occurrence by a factor of 75%. Finally, we highlight the use of receiver operating characteristic curves as a diagnostic for binary classifiers, and ultimately we use these to demonstrate the competitive predictive performance of GLMs against the popular technique of artificial neural networks.