def mc_cor(self,
mc_cov_mat=None,
method='inse',
adjust=False,
b=None,
r=3):
if mc_cov_mat is None:
return st.mc_cor(self.get_samples(),
method=method,
adjust=adjust,
b=b,
r=r,
rowvar=False)
else:
return st.cor_from_cov(mc_cov_mat)
def mc_cor(self,
mc_cov_mat=None,
method='inse',
adjust=False,
b=None,
r=3):
return np.stack([
st.mc_cor(self.get_chain(i, key='sample'),
method=method,
adjust=adjust,
b=b,
r=r,
rowvar=False)
if mc_cov_mat is None else st.cor_from_cov(mc_cov_mat[i])
for i in range(self.num_chains())
])
# Examples of empirical correlation matrix computation using cor_from_cov function
# %% Load packages
import numpy as np
from kanga.stats import cor_from_cov
# %% Read chains
chains = np.genfromtxt('chain01.csv', delimiter=',')
num_iters, num_pars = chains.shape
# %% Compute correlation matrix via numpy cor_from_cov function
np_cor_matrix = cor_from_cov(np.cov(chains, rowvar=False))
print('Correlation matrix based on cor_from_cov function:\n{}'.format(
np_cor_matrix))