Decoding the age-chemical structure of the Milky Way disk: An application of Copulas and Elicitable Maps
Code to reproduce the results and plots in Patil et al. (2023) is available in this repository. Additionally, we provide modules to compute copulas and elicitable maps of any given multivariate data.
The copula and conditional_correlation modules provide a user-friendly way to apply copulas and elicitable maps. The following example shows its usage:
import numpy as np
from copula import copula
from conditional_correlation import conditional_correlation
# Generate random two-dimensional data
data = [np.random.rand(1000), np.random.rand(1000)]
# Create copula object
model = copula(norm_data)
# Compute copula space data (uniform margins)
u1 = model.F[0](data[0])
u2 = model.F[1](data[1])
# Generate a variable upon with your data is conditional
z = [np.random.rand(1000)]
# conditional data (x, y | z)
x, y = data[0], data[1]
# Compute conditional correlation rho(x, y | z) using elicitable maps
cm = conditional_correlation([x, y, z])
cm.Estimate(n_iter=1_000, n_print=200)
# Sample continuous function rho(x, y | z)
cm.rho(np.random.rand(100))
Aarya A Patil, Jo Bovy, Sebastian Jaimungal, Neige Frankel, Henry W Leung, Decoding the age–chemical structure of the Milky Way disc: an application of copulas and elicitable maps, Monthly Notices of the Royal Astronomical Society, Volume 526, Issue 2, December 2023, Pages 1997–2016, https://doi.org/10.1093/mnras/stad2820