def test_ridge(data):
estimator = make_pipeline(
StandardScaler(),
MLPRegressor(hidden_layer_sizes=(50, 15,))
)
ttr.check_model(data, "MLPRegressor", estimator)
def test(data):
svr = make_pipeline(
# PolynomialFeatures(),
StandardScaler(),
SVR(kernel="rbf", C=25, gamma="scale"),
)
ttr.check_model(data, "SVR rbf", svr)
def test_linear_model(data):
ttr.check_model(
data,
"SGD",
make_pipeline(
# StandardScaler(),
PolynomialFeatures(),
SGDRegressor(max_iter=1000, tol=1e-3)))
def test_search_random_forest(data):
grid_params = {
"randomforestregressor__n_estimators": [20, 50, 100, 150, 200],
"randomforestregressor__criterion": ["mse"],
"randomforestregressor__min_samples_leaf": [2, 4, 8, 16, 32],
}
model = make_pipeline(
ColumnRemover(("timeStamp", )),
RandomForestRegressor(),
)
ttr.search(data, "Random Forest", model, grid_params)
def test_scan(data):
parameters = {
"svr__kernel": ["rbf"],
# "kernel": ["rbf", "poly", "sigmoid"],
"svr__C": [1, 10, 25, 30],
"svr__gamma": ["auto", 2. / 16, 4. / 16, 1.]
}
svr = make_pipeline(
ColumnRemover(("timeStamp",)),
StandardScaler(),
SVR(),
)
ttr.search(data, "SVR ", svr, parameters)
def test_bagging(data):
ttr.check_model(data, "Bagging", BaggingRegressor())
def test_gradient(data):
ttr.check_model(data, "Gradient boost", GradientBoostingRegressor())
def test_ada(data):
ttr.check_model(data, "Ada boost", AdaBoostRegressor())
def test_random_forest(data):
ttr.check_model(data, "Random Forest", RandomForestRegressor())
def main():
rf = make_pipeline(
StandardScaler(),
SVR(kernel="rbf", C=25),
)
ttr.check_model("SVR", rf)