def get_search_space():
config = ConfigSpace()
max_depth = GridHyperparameter(name='max_depth', values=[3, 5])
config.add_hyper([max_depth])
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
n_neighbors = GridHyperparameter(name='n_neighbors',
values=[3, 5, 7, 10])
config.add_hyper([n_neighbors])
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
n_estimator_list = GridHyperparameter(name='n_estimators',
values=[100, 300, 500])
config.add_hyper(n_estimator_list)
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
c_list = UniformHyperparameter(name="C", low=0.1, high=10, size=3)
# dual = CategoryHyperparameter(name="dual", categories=[True, False])
# grid = GridHyperparameter(name="C", values=[10, 100])
config.add_hyper([c_list])
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
fit_intercept = CategoryHyperparameter(name="fit_intercept",
categories=[True, False])
# grid = GridHyperparameter(name="C", values=[1, 2, 3])
config.add_hyper([fit_intercept])
# config.get_hypers()
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
n_estimators = GridHyperparameter(name='n_estimators',
values=[100, 300, 500])
learning_rate = UniformHyperparameter(name='learning_rate',
low=0.01,
high=1.0,
size=2)
config.add_hyper([n_estimators, learning_rate])
return config.get_hypers()
def get_search_space():
config = ConfigSpace()
# Let's try base estimator with DT, as in real world that also make improvement.
base_estimators = GridHyperparameter(name='base_estimator',
values=[None, DTRegressor()])
n_estimators = GridHyperparameter(name='n_estimators',
values=[50, 70, 100])
learning_rate = UniformHyperparameter(name='learning_rate',
low=0.01,
high=1.0,
size=2)
config.add_hyper([n_estimators, learning_rate])
return config.get_hypers()