def evaluate_new_feature(self, prev_subset: list, new_feature, X_f: dict,
X_t: dict, y: np.array) -> float:
A = prev_subset + [new_feature]
scores = list()
if self.n_jobs > 1:
scores = Parallel(n_jobs=self.n_jobs)(
delayed(self.score_function)(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=clone(self.decision_function))
for result in (split_dataset(X_t, X_f, y, self.seeds[i], 1 -
self.train_share)
for i in range(self.n_cv_ffs)))
else:
for i in range(self.n_cv_ffs):
result = split_dataset(X_t, X_f, y, self.seeds[i],
1 - self.train_share)
scores.append(
self.score_function(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=self.decision_function))
return float(np.mean(scores))
async def eval_new_feature(subset: list, new_feature: str, X_f: dict,
X_t: dict, y, n_cv_ffs: int, n_jobs: int,
seeds: list, train_share: float,
score_function_components, decision_function):
A = subset + [new_feature]
likelihoods = list()
if n_jobs > 1:
likelihoods = Parallel(n_jobs=n_jobs)(
delayed(score_function_components)(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=decision_function)
for result in (split_dataset(X_t, X_f, y, seeds[i], 1 -
train_share)
for i in range(n_cv_ffs)))
else:
for i in range(n_cv_ffs):
result = split_dataset(X_t, X_f, y, seeds[i], 1 - train_share)
likelihoods.append(
score_function_components(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=decision_function))
return {
k: float(np.mean([l[k] for l in likelihoods]))
for k in likelihoods[0].keys()
}
def evaluate_new_feature(self, prev_subset, new_feature, X_f, X_t,
y) -> dict:
A = prev_subset + [new_feature]
likelihoods = list()
if self.n_jobs > 1:
likelihoods = Parallel(n_jobs=self.n_jobs)(
delayed(self.score_function_components)(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=clone(self.decision_function))
for result in (split_dataset(X_t, X_f, y, self.seeds[i], 1 -
self.train_share)
for i in range(self.n_cv_ffs)))
else:
for i in range(self.n_cv_ffs):
result = split_dataset(X_t, X_f, y, self.seeds[i],
1 - self.train_share)
likelihoods.append(
self.score_function_components(
A=A,
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
decision_function=self.decision_function))
return {
k: np.mean([l[k] for l in likelihoods])
for k in likelihoods[0].keys()
}
def score(self, numeric_features: list) -> np.float128:
scores = list()
for i in range(self.n_cv):
result = split_dataset(self.X_t, self.X_f, self.y, self.seeds[i],
1 - self.train_share)
scores.append(
self._score_function(
A=[self.feature_list[f] for f in numeric_features],
X_f=result['train']['transformed'],
X_f_test=result['test']['transformed'],
X_t=result['train']['plain'],
X_t_test=result['test']['plain'],
y=result['train']['target'],
y_test=result['test']['target'],
))
return np.mean(scores)
def evaluate(self, dataset: HoldoutDataset, seed):
result = split_dataset(
dataset.get_features(False),
dataset.get_features(self.requires_linearisation),
dataset.get_target(), seed, int(1))
subset = self.optimizer.select(result['train']['plain'],
result['train']['transformed'],
result['train']['target'],
dataset.get_continuous_feature_names())
y_pred = self.train_and_fit(subset, result['train']['transformed'],
result['train']['target'],
result['test']['transformed'])
return {
'target': np.squeeze(result['test']['target']),
'pred': y_pred,
'subset': subset
}