def correlations(signal,
signal_weight,
background,
background_weight,
fields,
category,
output_suffix=''):
names = [
VARIABLES[field]['title'] if field in VARIABLES else field
for field in fields
]
# draw correlation plots
plot_corrcoef_matrix(signal,
fields=names,
output_name=os.path.join(
PLOTS_DIR, "correlation_signal_%s%s.png" %
(category.name, output_suffix)),
title='%s Signal' % category.label,
weights=signal_weight)
plot_corrcoef_matrix(background,
fields=names,
output_name=os.path.join(
PLOTS_DIR, "correlation_background_%s%s.png" %
(category.name, output_suffix)),
title='%s Background' % category.label,
weights=background_weight)
def correlations(signal, signal_weight, background, background_weight, fields, category, output_suffix=""):
names = [VARIABLES[field]["title"] if field in VARIABLES else field for field in fields]
# draw correlation plots
plot_corrcoef_matrix(
signal,
fields=names,
output_name=os.path.join(PLOTS_DIR, "correlation_signal_%s%s.png" % (category.name, output_suffix)),
title="%s Signal" % category.label,
weights=signal_weight,
)
plot_corrcoef_matrix(
background,
fields=names,
output_name=os.path.join(PLOTS_DIR, "correlation_background_%s%s.png" % (category.name, output_suffix)),
title="%s Background" % category.label,
weights=background_weight,
)
import string
from rootpy.plotting.contrib import plot_corrcoef_matrix
if __name__ == '__main__':
import numpy as np
n_vars = 10
var_names = ['var_%s' % s for s in string.lowercase[:n_vars]]
def random_symm(n):
a = np.random.random_integers(-10, 10, size=(n, n))
return (a + a.T) / 2
data = np.random.multivariate_normal(-np.random.random(n_vars) * 3,
cov=random_symm(n_vars),
size=100000)
weights = np.random.randint(1, 10, 100000)
plot_corrcoef_matrix(data,
var_names,
'correlations.png',
weights=weights,
title='correlations')
import string
from rootpy.plotting.contrib import plot_corrcoef_matrix
if __name__ == '__main__':
import numpy as np
n_vars = 10
var_names = ['var_%s' % s for s in string.lowercase[:n_vars]]
def random_symm(n):
a = np.random.random_integers(-10, 10, size=(n, n))
return (a + a.T) / 2
data = np.random.multivariate_normal(
-np.random.random(n_vars) * 3, cov=random_symm(n_vars), size=100000)
weights = np.random.randint(1, 10, 100000)
plot_corrcoef_matrix(
data, var_names, 'correlations.png',
weights=weights, title='correlations')
