def fit(self, x, y=None):
if y is not None:
xdot = y
else:
xdot = self.derivative.transform(x)
if self.operators is not None:
feature_transformer = SymbolicFeatures(exponents=np.linspace(
1, self.degree, self.degree),
operators=self.operators)
else:
feature_transformer = PolynomialFeatures(degree=self.degree,
include_bias=False)
steps = [
("features", feature_transformer),
("model",
STRidge(alpha=self.alpha, threshold=self.threshold, **self.kw)),
]
self.model = MultiOutputRegressor(Pipeline(steps), n_jobs=self.n_jobs)
self.model.fit(x, xdot)
self.n_input_features_ = self.model.estimators_[0].steps[0][
1].n_input_features_
self.n_output_features_ = self.model.estimators_[0].steps[0][
1].n_output_features_
return self
def __init__(self, strategy, **kw):
self.strategy = strategy
self.kw = kw
super().__init__(steps=[
("selection", ColumnSelector(index=self.strategy.index)),
("features", SymbolicFeatures(exponents=self.strategy.exponents,
operators=self.strategy.operators,
consider_products=self.strategy.consider_products)),
("regression", strategy.base(warm_start=True, **self.kw))])
from sklearn.metrics import make_scorer
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import train_test_split
from sklearn.utils.validation import check_random_state
from sparsereg.model.base import print_model
from sparsereg.model.group_lasso import SparseGroupLasso
from sparsereg.preprocessing.symfeat import SymbolicFeatures
rng = check_random_state(42)
x = rng.normal(size=(10000, 1))
y = np.cos(x[:, 0]) + x[:, 0]**2 + x[:, 0]**3 # + 0.01*rng.normal(size=1000)
x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=rng)
pre = SymbolicFeatures(exponents=[1, 2],
operators={
"sin": np.sin,
"cos": np.cos
}).fit(x_train)
features_train = pre.transform(x_train)
features_test = pre.transform(x_test)
km = AgglomerativeClustering(n_clusters=4).fit(features_train.T)
labels = defaultdict(list)
for k, v in zip(pre.get_feature_names(), km.labels_):
labels[v].append(k)
print(labels)
params = {"alpha": [0.001, 0.01, 0.02, 0.05], "normalize": [True]}
scorer = make_scorer(explained_variance_score)
sgl = SparseGroupLasso(groups=km.labels_, rho=0.3, alpha=0.02)
l = Lasso()
for model in [sgl, l]:
grid = GridSearchCV(model, params, n_jobs=1, scoring=scorer,