def compute_theil_on_bins(y_pred, mask, bin_indices, target_efficiencies,
sample_weight):
y_pred = column_or_1d(y_pred)
sample_weight = check_sample_weight(y_pred, sample_weight=sample_weight)
# ignoring events from other classes
y_pred = y_pred[mask]
bin_indices = bin_indices[mask]
sample_weight = sample_weight[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices,
sample_weight=sample_weight)
cuts = compute_cut_for_efficiency(target_efficiencies,
mask=numpy.ones(len(y_pred), dtype=bool),
y_pred=y_pred,
sample_weight=sample_weight)
result = 0.
for cut in cuts:
bin_efficiencies = compute_bin_efficiencies(
y_pred,
bin_indices=bin_indices,
cut=cut,
sample_weight=sample_weight)
result += theil(bin_efficiencies, weights=bin_weights)
return result / len(cuts)
def compute_sde_on_bins(y_pred,
mask,
bin_indices,
target_efficiencies,
power=2.,
sample_weight=None):
# ignoring events from other classes
sample_weight = check_sample_weight(y_pred, sample_weight=sample_weight)
y_pred = y_pred[mask]
bin_indices = bin_indices[mask]
sample_weight = sample_weight[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices,
sample_weight=sample_weight)
cuts = compute_cut_for_efficiency(target_efficiencies,
mask=numpy.ones(len(y_pred), dtype=bool),
y_pred=y_pred,
sample_weight=sample_weight)
result = 0.
for cut in cuts:
bin_efficiencies = compute_bin_efficiencies(
y_pred,
bin_indices=bin_indices,
cut=cut,
sample_weight=sample_weight)
result += weighted_deviation(bin_efficiencies,
weights=bin_weights,
power=power)
return (result / len(cuts))**(1. / power)
def compute_theil_on_bins(y_pred, mask, bin_indices, target_efficiencies, sample_weight):
y_pred = column_or_1d(y_pred)
sample_weight = check_sample_weight(y_pred, sample_weight=sample_weight)
# ignoring events from other classes
y_pred = y_pred[mask]
bin_indices = bin_indices[mask]
sample_weight = sample_weight[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices, sample_weight=sample_weight)
cuts = compute_cut_for_efficiency(target_efficiencies, mask=numpy.ones(len(y_pred), dtype=bool),
y_pred=y_pred, sample_weight=sample_weight)
result = 0.
for cut in cuts:
bin_efficiencies = compute_bin_efficiencies(y_pred, bin_indices=bin_indices,
cut=cut, sample_weight=sample_weight)
result += theil(bin_efficiencies, weights=bin_weights)
return result / len(cuts)
def compute_sde_on_bins(y_pred, mask, bin_indices, target_efficiencies, power=2., sample_weight=None):
# ignoring events from other classes
sample_weight = check_sample_weight(y_pred, sample_weight=sample_weight)
y_pred = y_pred[mask]
bin_indices = bin_indices[mask]
sample_weight = sample_weight[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices, sample_weight=sample_weight)
cuts = compute_cut_for_efficiency(target_efficiencies, mask=numpy.ones(len(y_pred), dtype=bool),
y_pred=y_pred, sample_weight=sample_weight)
result = 0.
for cut in cuts:
bin_efficiencies = compute_bin_efficiencies(y_pred, bin_indices=bin_indices,
cut=cut, sample_weight=sample_weight)
result += weighted_deviation(bin_efficiencies, weights=bin_weights, power=power)
return (result / len(cuts)) ** (1. / power)
def bin_based_cvm(y_pred, sample_weight, bin_indices):
"""Cramer-von Mises similarity, quite slow meanwhile"""
assert len(y_pred) == len(sample_weight) == len(bin_indices)
bin_weights = compute_bin_weights(bin_indices=bin_indices, sample_weight=sample_weight)
result = 0.
global_data, global_weight, global_F = prepare_distribution(y_pred, weights=sample_weight)
for bin, bin_weight in enumerate(bin_weights):
if bin_weight <= 0:
continue
bin_mask = bin_indices == bin
local_distribution = y_pred[bin_mask]
local_weights = sample_weight[bin_mask]
result += bin_weight * _cvm_2samp_fast(global_data, local_distribution,
global_weight, local_weights, global_F)
return result
def bin_based_ks(y_pred, mask, sample_weight, bin_indices):
"""Kolmogorov-Smirnov flatness on bins"""
assert len(y_pred) == len(sample_weight) == len(bin_indices) == len(mask)
y_pred = y_pred[mask]
sample_weight = sample_weight[mask]
bin_indices = bin_indices[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices, sample_weight=sample_weight)
prepared_data, prepared_weight, prep_F = prepare_distribution(y_pred, weights=sample_weight)
result = 0.
for bin, bin_weight in enumerate(bin_weights):
if bin_weight <= 0:
continue
local_distribution = y_pred[bin_indices == bin]
local_weights = sample_weight[bin_indices == bin]
result += bin_weight * \
_ks_2samp_fast(prepared_data, local_distribution, prepared_weight, local_weights, prep_F)
return result
def bin_based_cvm(y_pred, sample_weight, bin_indices):
"""Cramer-von Mises similarity, quite slow meanwhile"""
assert len(y_pred) == len(sample_weight) == len(bin_indices)
bin_weights = compute_bin_weights(bin_indices=bin_indices,
sample_weight=sample_weight)
result = 0.
global_data, global_weight, global_F = prepare_distribution(
y_pred, weights=sample_weight)
for bin, bin_weight in enumerate(bin_weights):
if bin_weight <= 0:
continue
bin_mask = bin_indices == bin
local_distribution = y_pred[bin_mask]
local_weights = sample_weight[bin_mask]
result += bin_weight * _cvm_2samp_fast(global_data, local_distribution,
global_weight, local_weights,
global_F)
return result
def bin_based_ks(y_pred, mask, sample_weight, bin_indices):
"""Kolmogorov-Smirnov flatness on bins"""
assert len(y_pred) == len(sample_weight) == len(bin_indices) == len(mask)
y_pred = y_pred[mask]
sample_weight = sample_weight[mask]
bin_indices = bin_indices[mask]
bin_weights = compute_bin_weights(bin_indices=bin_indices,
sample_weight=sample_weight)
prepared_data, prepared_weight, prep_F = prepare_distribution(
y_pred, weights=sample_weight)
result = 0.
for bin, bin_weight in enumerate(bin_weights):
if bin_weight <= 0:
continue
local_distribution = y_pred[bin_indices == bin]
local_weights = sample_weight[bin_indices == bin]
result += bin_weight * \
_ks_2samp_fast(prepared_data, local_distribution, prepared_weight, local_weights, prep_F)
return result