Explore an entirely analytic model which includes no Poisson error

In this example, we explore the usefulness of a completely analytic solution to an ideal “sample”. For details, see Appendix D.3.3 of the MRP paper. In brief, the “ideal sample” is composed of non-Poisson limited haloes extracted from a pure MRP distribution within some physical volume. As such, the solution is a priori the input parameters of the MRP distribution. Finding the covariance of those parameters is our task.

The whole framework of this problem has already been implemented in mrpy.extra.analytic_model. We will use that framework to answer some questions.

In [1]:
# Imports
%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np

from mrpy.extra.analytic_model import IdealAnalytic
In [2]:
# Fiducial Parameters
beta = 0.75
alpha = -1.85
hs = 14.5
V=400.0**3 # Physical volume

# Some constants
log_mmin = np.linspace(11,15,10)

What is the expected variance of parameters versustruncation mass?

In [3]:
fig,ax = plt.subplots(3,1,sharex=True,sharey=False,figsize=(6,7),subplot_kw={"yscale":'log'},
                     gridspec_kw={"hspace":0.0})

K = IdealAnalytic(log_mmin=log_mmin,logHs=hs,alpha=alpha,beta=beta,lnA=0.0)
ax[0].plot(log_mmin-hs,np.sqrt(K.cov[:,0,0])/np.abs(hs),lw=2)
ax[1].plot(log_mmin-hs,np.sqrt(K.cov[:,1,1])/np.abs(alpha),lw=2)
ax[2].plot(log_mmin-hs,np.sqrt(K.cov[:,2,2])/np.abs(beta),lw=2)

#ax[0].legend(loc=0,ncol=2)

ax[0].set_ylabel(r"$\sigma_{\log \mathcal{H}_\star}/|\log \mathcal{H}_\star|$",fontsize=15)
ax[1].set_ylabel(r"$\sigma_\alpha/|\alpha|$",fontsize=15)
ax[2].set_ylabel(r"$\sigma_\beta/|\beta|$",fontsize=15)
ax[2].set_xlabel(r"$\log_{10}\left(m_{min}/\mathcal{H}_\star\right)$",fontsize=15)
Out[3]:
<matplotlib.text.Text at 0x7f5681ccfb10>
../_images/examples_explore_analytic_model_5_1.png

What is the expected correlation of parameters versus truncation mass?

In [4]:
fig,ax = plt.subplots(3,1,sharex=True,sharey=False,figsize=(6,7),
                     gridspec_kw={"hspace":0.1})

K = IdealAnalytic(log_mmin=log_mmin,logHs=hs,alpha=alpha,beta=beta,lnA=0.0)
ax[0].plot(log_mmin-hs,K.corr[:,0,1],lw=2)
ax[1].plot(log_mmin-hs,K.corr[:,1,2],lw=2)
ax[2].plot(log_mmin-hs,K.corr[:,0,2],lw=2)

#ax[0].legend(loc=0,ncol=2)

ax[0].set_ylabel(r"$\log \mathcal{H}_\star-\alpha$",fontsize=15)
ax[1].set_ylabel(r"$\alpha-\beta$",fontsize=15)
ax[2].set_ylabel(r"$\log \mathcal{H}_\star-\beta$",fontsize=15)
ax[2].set_xlabel(r"$\log_{10}\left(m_{min}/\mathcal{H}_\star\right)$",fontsize=15)
Out[4]:
<matplotlib.text.Text at 0x7f5681b27090>
../_images/examples_explore_analytic_model_7_1.png