def get_dataset_dictionaries(self, lengths=[5, 6]):
x_train = dict()
y_train = dict()
x_test = dict()
y_test = dict()
for n_obj in lengths:
self.kwargs["n_objects"] = n_obj
seed = self.random_state.randint(2 ** 32, dtype="uint32")
total_instances = self.n_test_instances + self.n_train_instances
self.kwargs["n_instances"] = total_instances
X, Y = self.dataset_function(**self.kwargs, seed=seed)
x_1, x_2, y_1, y_2 = train_test_split(
X, Y, random_state=self.random_state, test_size=self.n_test_instances
)
if self.standardize:
x_1, x_2 = standardize_features(x_1, x_2)
x_train[n_obj], x_test[n_obj], y_train[n_obj], y_test[n_obj] = (
x_1,
x_2,
y_1,
y_2,
)
logger.info("Done")
return x_train, y_train, x_test, y_test
def get_single_train_test_split(self):
self.dataset_function()
x_train, x_test, y_train, y_test = train_test_split(self.X, self.Y, random_state=self.random_state,
test_size=self.n_test_instances)
if self.standardize:
x_train, x_test = standardize_features(x_train, x_test)
self.logger.info('Done')
return x_train, y_train, x_test, y_test
def splitter(self, iter):
for train_idx, test_idx in iter:
x_train, y_train, x_test, y_test = (
self.X[train_idx],
self.Y[train_idx],
self.X[test_idx],
self.Y[test_idx],
)
x_train, x_test = standardize_features(x_train, x_test)
yield x_train, y_train, x_test, y_test
def splitter(self, iter):
for i in iter:
seed = self.random_state.randint(2 ** 32, dtype="uint32")
total_instances = self.n_test_instances + self.n_train_instances
self.kwargs["n_instances"] = total_instances
X, Y = self.dataset_function(**self.kwargs, seed=seed)
x_train, x_test, y_train, y_test = train_test_split(
X, Y, random_state=self.random_state, test_size=self.n_test_instances
)
if self.standardize:
x_train, x_test = standardize_features(x_train, x_test)
yield x_train, y_train, x_test, y_test
def get_dataset_dictionaries(self, lengths=[5, 6]):
x_train = dict()
y_train = dict()
x_test = dict()
y_test = dict()
for n_obj in lengths:
self.n_objects = n_obj
self.dataset_function()
x_1, x_2, y_1, y_2 = train_test_split(self.X, self.Y, random_state=self.random_state,
test_size=self.n_test_instances)
if self.standardize:
x_1, x_2 = standardize_features(x_1, x_2)
x_train[n_obj], x_test[n_obj], y_train[n_obj], y_test[n_obj] = x_1, x_2, y_1, y_2
self.logger.info('Done')
return x_train, y_train, x_test, y_test
def get_single_train_test_split(self):
seed = self.random_state.randint(2 ** 32, dtype="uint32")
total_instances = self.n_test_instances + self.n_train_instances
self.kwargs["n_instances"] = total_instances
self.X, self.Y = self.dataset_function(**self.kwargs, seed=seed)
self.__check_dataset_validity__()
x_train, x_test, y_train, y_test = train_test_split(
self.X,
self.Y,
random_state=self.random_state,
test_size=self.n_test_instances,
)
if self.standardize:
x_train, x_test = standardize_features(x_train, x_test)
logger.info("Done")
return x_train, y_train, x_test, y_test
def get_single_train_test_split(self):
splits = dict()
cv_iter = ShuffleSplit(n_splits=1,
random_state=self.random_state,
test_size=0.80)
for n_obj, arr in self.X_dict.items():
if arr.shape[0] == 1:
splits[n_obj] = ([0], [0])
else:
splits[n_obj] = list(cv_iter.split(arr))[0]
self.X_train = dict()
self.Y_train = dict()
self.X_test = dict()
self.Y_test = dict()
for n_obj, itr in splits.items():
train_idx, test_idx = itr
self.X_train[n_obj] = np.copy(self.X_dict[n_obj][train_idx])
self.X_test[n_obj] = np.copy(self.X_dict[n_obj][test_idx])
self.Y_train[n_obj] = np.copy(self.Y_dict[n_obj][train_idx])
self.Y_test[n_obj] = np.copy(self.Y_dict[n_obj][test_idx])
self.X, self.Y = self.sub_sampling_from_dictionary()
self.__check_dataset_validity__()
self.X, self.X_test = standardize_features(self.X, self.X_test)
return self.X, self.Y, self.X_test, self.Y_test
def get_single_train_test_split(self):
self.X, self.Y = self.sub_sampling_from_dictionary(train_test="train")
self.__check_dataset_validity__()
self.X, self.X_test = standardize_features(self.X, self.X_test)
return self.X, self.Y, self.X_test, self.Y_test
def get_dataset_dictionaries(self):
self.X_test, self.X_test = standardize_features(self.X, self.X_test)
return self.X_train, self.Y_train, self.X_test, self.Y_test
def get_single_train_test_split(self):
X_train, Y_train = self.sub_sampling_from_dictionary(train_test="train")
X_test, Y_test = self.sub_sampling_from_dictionary(train_test="test")
X_train, X_test = standardize_features(X_train, X_test)
return X_train, Y_train, X_test, Y_test
