def bandwidth_kdeCV(X,
kde=None,
bw=10**np.linspace(-2, 1),
cv=1,
verbose=True,
kernel='gaussian',
metric='euclidean',
atol=0,
leaf_size=40):
"""
Estimate the bandwidth using cross validation.
"""
if kde is None:
kde = KernelDensity(kernel=kernel,
metric=metric,
atol=atol,
leaf_size=leaf_size)
Jh = np.array([])
for i, h in enumerate(bw):
kde.bandwidth = h
kde.fit(X)
log_pdf = kde.score_samples(X)
pdf = np.exp(log_pdf)
jh_ = cross_val_loss_kdeCV_loo(
X, kde, fx=pdf) if cv == 1 else cross_val_loss_kdeCV_kFold(
X, kde, fx=pdf, n_splits=cv)
Jh = np.append(Jh, jh_)
if verbose:
print('Completed: {0:.2f} %'.format(100 * (i + 1) / bw.size))
return bw[Jh.argmin()]
