def test_nested_model_stack(teardown):
x_data = iris.data
y_t_data = iris.target
random_state = 123
n_components = 2
# ----------- baikal way
stacked_model_baikal = make_naive_stacked_model(
n_components, random_state, x_data, y_t_data
)
y_pred_baikal = stacked_model_baikal.predict(x_data)
# ----------- traditional way
# Submodel 1
submodel1 = LogisticRegression(
multi_class="multinomial", solver="lbfgs", random_state=random_state
)
pca = PCA(n_components=n_components, random_state=random_state)
pca.fit(x_data)
pca_trans = pca.transform(x_data)
submodel1.fit(pca_trans, y_t_data)
submodel1_pred = submodel1.predict(pca_trans)
# Submodel 2 (a nested stacked model)
random_forest = RandomForestClassifier(random_state=random_state)
random_forest.fit(x_data, y_t_data)
random_forest_pred = random_forest.predict(x_data)
extra_trees = ExtraTreesClassifier(random_state=random_state)
extra_trees.fit(x_data, y_t_data)
extra_trees_pred = extra_trees.predict(x_data)
features = np.stack([random_forest_pred, extra_trees_pred], axis=1)
submodel2 = LogisticRegression(
multi_class="multinomial", solver="lbfgs", random_state=random_state
)
submodel2.fit(features, y_t_data)
submodel2_pred = submodel2.predict(features)
# Stacked model
features = np.stack([submodel1_pred, submodel2_pred], axis=1)
stacked_model_traditional = LogisticRegression(
multi_class="multinomial", solver="lbfgs", random_state=random_state
)
stacked_model_traditional.fit(features, y_t_data)
y_pred_traditional = stacked_model_traditional.predict(features)
assert_array_equal(y_pred_baikal, y_pred_traditional)
def make_naive_stacked_model(n_components, random_state, x_data, y_t_data):
# An unnecessarily complex Model
# Sub-model 1
x1 = Input(name="x1")
y1_t = Input(name="y1_t")
h1 = PCA(n_components=n_components, random_state=random_state, name="pca_sub1")(x1)
y1 = LogisticRegression(
multi_class="multinomial",
solver="lbfgs",
random_state=random_state,
name="logreg_sub1",
)(h1, y1_t)
submodel1 = Model(x1, y1, y1_t, name="submodel1")
# Sub-model 2 (a nested stacked model)
x2 = Input(name="x2")
y2_t = Input(name="y2_t")
y2_1 = RandomForestClassifier(random_state=random_state, name="rforest_sub2")(
x2, y2_t
)
y2_2 = ExtraTreesClassifier(random_state=random_state, name="extrees_sub2")(
x2, y2_t
)
features = Stack(axis=1, name="stack_sub2")([y2_1, y2_2])
y2 = LogisticRegression(
multi_class="multinomial",
solver="lbfgs",
random_state=random_state,
name="logreg_sub2",
)(features, y2_t)
submodel2 = Model(x2, y2, y2_t, name="submodel2")
# Stack of submodels
x = Input(name="x")
y_t = Input(name="y_t")
y1 = submodel1(x, y_t)
y2 = submodel2(x, y_t)
features = Stack(axis=1, name="stack")([y1, y2])
y = LogisticRegression(
multi_class="multinomial",
solver="lbfgs",
random_state=random_state,
name="logreg_stacked",
)(features, y_t)
stacked_model_baikal = Model(x, y, y_t, name="stacked")
stacked_model_baikal.fit(x_data, y_t_data)
return stacked_model_baikal