def __init__(self, pipeline_id, do_prep='True', prep_dict=None):
self.pid = pipeline_id
self.do_prep = do_prep == 'True'
self.cat_idx = None
self.float_idx = None
self.prep_dict = prep_dict
BaseHelper.__init__(self)
def __init__(self,
pipeline_id=None,
do_prep='True',
prep_dict={'impute_strategy': 'impute_knn5'},
gridpoints=4,
inner_cv=None,
groupcount=None,
impute_strategy=None,
float_idx=None,
cat_idx=None,
bestT=False,
cv_splits=5,
cv_repeats=2):
self.pipeline_id = pipeline_id
self.do_prep = do_prep == 'True' if type(do_prep) != bool else do_prep
self.gridpoints = gridpoints
self.inner_cv = inner_cv
self.groupcount = groupcount
self.bestT = bestT
self.cat_idx = cat_idx
self.float_idx = float_idx
self.prep_dict = prep_dict
self.cv_splits = cv_splits
self.cv_repeats = cv_repeats
self.impute_strategy = impute_strategy
if impute_strategy:
self.prep_dict["impute_strategy"] = self.impute_strategy
BaseHelper.__init__(self)
def __init__(self,
pipeline_id=None,
do_prep='True',
functional_form_search="False",
prep_dict={'impute_strategy': 'impute_knn5'},
gridpoints=4,
inner_cv=None,
groupcount=None,
bestT=False,
cat_idx=None,
float_idx=None,
flex_kwargs={},
cv_splits=5,
cv_repeats=2):
self.pipeline_id = pipeline_id
self.do_prep = do_prep == 'True' if type(do_prep) != bool else do_prep
self.functional_form_search = functional_form_search if type(
functional_form_search
) == bool else functional_form_search == "True"
self.gridpoints = gridpoints
self.inner_cv = inner_cv
self.groupcount = groupcount
self.bestT = bestT
self.cat_idx = cat_idx
self.float_idx = float_idx
self.prep_dict = prep_dict
self.flex_kwargs = flex_kwargs
self.cv_splits = cv_splits
self.cv_repeats = cv_repeats
self.flex_kwargs["robust"] = True
BaseHelper.__init__(self)
def __init__(self,
pipeline_id=None,
do_prep=True,
prep_dict={'impute_strategy': 'impute_knn5'},
inner_cv=None,
bestT=False,
cat_idx=None,
float_idx=None,
est_kwargs=None,
cv_splits=5,
cv_repeats=2):
self.pipeline_id = pipeline_id
self.do_prep = do_prep == 'True' if type(do_prep) != bool else do_prep
self.bestT = bestT
self.cat_idx = cat_idx
self.float_idx = float_idx
self.inner_cv = inner_cv
self.prep_dict = prep_dict
self.est_kwargs = est_kwargs
self.cv_splits = cv_splits
self.cv_repeats = cv_repeats
# self.impute_strategy = impute_strategy
# if impute_strategy:
# self.prep_dict["impute_strategy"] = self.impute_strategy
BaseHelper.__init__(self)
def __init__(self, pipeline_id=None, do_prep='True', prep_dict=None):
self.pipeline_id = pipeline_id
self.do_prep = do_prep == 'True' if type(do_prep) != bool else do_prep
self.cat_idx = None
self.float_idx = None
self.prep_dict = prep_dict
BaseHelper.__init__(self)
def __init__(self,
pipeline_id,
do_prep='True',
prep_dict=None,
impute_strategy=None,
gridpoints=4,
inner_cv=None,
groupcount=None,
float_idx=None,
cat_idx=None,
bestT=False):
self.pid = pipeline_id
self.do_prep = do_prep == 'True'
self.gridpoints = gridpoints
self.groupcount = groupcount
self.float_idx = float_idx
self.cat_idx = cat_idx
self.bestT = bestT
self.inner_cv = inner_cv
self.prep_dict = {
'impute_strategy':
self.hyper_parameters["impute_strategy"]["value"]
} if prep_dict is None else prep_dict
if impute_strategy:
self.prep_dict["impute_strategy"] = impute_strategy
self.flags = None
BaseHelper.__init__(self)
def __init__(self,
pipeline_id,
do_prep='True',
prep_dict={'impute_strategy': 'impute_knn5'},
gridpoints=4,
inner_cv=None,
groupcount=None,
impute_strategy=None,
bestT=False,
cat_idx=None,
float_idx=None):
self.pid = pipeline_id
self.do_prep = do_prep == 'True'
self.gridpoints = gridpoints
self.inner_cv = inner_cv
self.groupcount = groupcount
self.bestT = bestT
self.cat_idx = cat_idx
self.float_idx = float_idx
self.prep_dict = prep_dict
if impute_strategy:
self.prep_dict["impute_strategy"] = impute_strategy
BaseHelper.__init__(self)
def __init__(self,
pipeline_id,
do_prep='True',
functional_form_search=False,
prep_dict={'impute_strategy': 'impute_knn5'},
gridpoints=4,
inner_cv=None,
groupcount=None,
bestT=False,
cat_idx=None,
float_idx=None,
flex_kwargs={}):
self.pid = pipeline_id
self.do_prep = do_prep == 'True'
self.functional_form_search = functional_form_search
self.gridpoints = gridpoints
self.inner_cv = inner_cv
self.groupcount = groupcount
self.bestT = bestT
self.cat_idx = cat_idx
self.float_idx = float_idx
self.prep_dict = prep_dict
self.flex_kwargs = flex_kwargs
BaseHelper.__init__(self)
def predict(self, features=None):
train_idx_list, test_idx_list = zip(*list(self.get_cv().split(self.x, self.y)))
n, k = self.x.shape
data_idx = np.arange(n)
yhat_dict = {}
scores = {}
for idx, (estimator_name, result) in enumerate(self.cv_results.items()):
scores[estimator_name] = []
yhat_dict[estimator_name] = []
for r in range(self.cv_reps):
yhat = np.empty([n, ])
rows = []
for s in range(self.cv_folds): # s for split
m = r*self.cv_folds+s
cv_est = result['estimator'][m]
test_rows = test_idx_list[m]
yhat[test_rows] = cv_est.predict(self.x.iloc[test_rows])
rows.append(BaseHelper.score(self.y[test_rows], yhat[test_rows]))
scores[estimator_name].append(rows)
yhat_dict[estimator_name].append(yhat)
self.cv_yhat_dict = yhat_dict
return yhat_dict, scores