def test_generic_smoother():
x, y, poly = multivariate_sample_data()
alphas = [0.4, 0.7]
weights = [1, 1] # noqa: F841
gs = GenericSmoothers(poly.x, poly.smoothers)
gam_gs = GLMGam(y, smoother=gs, alpha=alphas)
gam_gs_res = gam_gs.fit()
gam_poly = GLMGam(y, smoother=poly, alpha=alphas)
gam_poly_res = gam_poly.fit()
assert_allclose(gam_gs_res.params, gam_poly_res.params)
def _split_train_test_smoothers(x, smoother, train_index, test_index):
"""split smoothers in test and train sets and create GenericSmoothers
Note: this does not take exog_linear into account
"""
train_smoothers = []
test_smoothers = []
for smoother in smoother.smoothers:
train_basis = smoother.basis[train_index]
train_der_basis = smoother.der_basis[train_index]
train_der2_basis = smoother.der2_basis[train_index]
train_cov_der2 = smoother.cov_der2
# TODO: Double check this part. cov_der2 is calculated with all data
train_x = smoother.x[train_index]
train_smoothers.append(
UnivariateGenericSmoother(train_x, train_basis, train_der_basis,
train_der2_basis, train_cov_der2,
smoother.variable_name + ' train'))
test_basis = smoother.basis[test_index]
test_der_basis = smoother.der_basis[test_index]
test_der2_basis = smoother.der2_basis[test_index]
test_cov_der2 = smoother.cov_der2
# TODO: Double check this part. cov_der2 is calculated with all data
test_x = smoother.x[test_index]
test_smoothers.append(
UnivariateGenericSmoother(test_x, test_basis, test_der_basis,
train_der2_basis, test_cov_der2,
smoother.variable_name + ' test'))
train_multivariate_smoothers = GenericSmoothers(x[train_index],
train_smoothers)
test_multivariate_smoothers = GenericSmoothers(x[test_index],
test_smoothers)
return train_multivariate_smoothers, test_multivariate_smoothers