def getModel(self, _params):
return KNeighborsClassifier(
n_neighbors=_params['n_neighbors'],
weights=_params['weights'],
algorithm=_params['algorithm'],
leaf_size=_params['leaf_size'],
p=_params['p'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return Perceptron(
penalty=_params['penalty'],
alpha=_params['alpha'],
fit_intercept=_params['fit_intercept'],
max_iter=int(_params['max_iter']),
shuffle=_params['shuffle'],
early_stopping=_params['early_stopping'],
validation_fraction=_params['validation_fraction'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
estimator_list = []
for idx, job in enumerate(self.cantidate_job_list):
if idx == _params['max_estimator']:
break
estimator_list.append((job.getJobName(), job.trained_model))
return Voting(
estimators=estimator_list,
# verbosity = 0,
n_jobs= definitions.getNumberOfCore(),
)
def getModel(self, _params):
return LogisticRegression(
penalty=_params['penalty'],
dual=_params['dual'],
C=_params['C'],
fit_intercept=_params['fit_intercept'],
intercept_scaling=_params['intercept_scaling'],
solver=_params['solver'],
max_iter=int(_params['max_iter']),
l1_ratio=_params['l1_ratio'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return XGBRegressor(
max_depth=int(_params['max_depth']),
booster=_params['booster'],
gamma=_params['gamma'],
min_child_weight=int(_params['min_child_weight']),
subsample=_params['subsample'],
colsample_bytree=_params['colsample_bytree'],
eta=_params['eta'],
verbosity=0,
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return LGBMRegressor(
max_depth= int(_params['max_depth']),
num_leaves= int((2 ** np.maximum(np.minimum(_params['max_depth'], 12), 2)) * 0.6),
boosting_type= _params['boosting_type'],
learning_rate= _params['learning_rate'],
n_estimators= int(_params['n_estimators']),
bagging_fraction= _params['bagging_fraction'],
bagging_freq= _params['bagging_freq'],
feature_fraction= _params['feature_fraction'],
max_bin = int(_params['max_bin']),
min_data_in_leaf = int(_params['min_data_in_leaf']),
n_jobs= definitions.getNumberOfCore(),
)
def getModel(self, _params):
return SGDClassifier(
penalty=_params['penalty'],
alpha=_params['alpha'],
l1_ratio=_params['l1_ratio'],
fit_intercept=bool(_params['fit_intercept']),
max_iter=int(_params['max_iter']),
shuffle=bool(_params['shuffle']),
learning_rate=_params['learning_rate'],
eta0=_params['eta0'],
early_stopping=bool(_params['early_stopping']),
class_weight=_params['class_weight'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return MLPRegressor(
hidden_layer_sizes=(int(_params['layer_1']),
int(_params['layer_2']),
int(_params['layer_3'])),
activation=_params['activation'],
solver=_params['solver'],
alpha=_params['alpha'],
learning_rate=_params['learning_rate'],
max_iter=int(_params['max_iter']),
validation_fraction=_params['validation_fraction'],
beta_1=_params['beta_1'],
beta_2=_params['beta_2'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return RandomForest(
n_estimators= int(_params['n_estimators']),
# criterion= _params['criterion'],
max_depth= _params['max_depth'],
# min_samples_split= _params['min_samples_split'],
# min_samples_leaf= _params['min_samples_leaf'],
min_weight_fraction_leaf= _params['min_weight_fraction_leaf'],
max_features= _params['max_features'],
# max_leaf_nodes= int(_params['max_leaf_nodes']),
# min_impurity_decrease= _params['min_impurity_decrease'],
# bootstrap= _params['bootstrap'],
oob_score= _params['oob_score'],
# ccp_alpha= _params['ccp_alpha'],
n_jobs= definitions.getNumberOfCore(),
)
def getModel(self, _params):
return LGBMClassifier(
boosting_type=_params['boosting_type'],
num_leaves=int(
(2**np.maximum(np.minimum(_params['max_depth'], 12), 2)) *
0.6),
max_depth=int(_params['max_depth']),
learning_rate=_params['learning_rate'],
n_estimators=int(_params['n_estimators']),
subsample_for_bin=int(_params['subsample_for_bin']),
min_split_gain=_params['min_split_gain'],
min_child_weight=_params['min_child_weight'],
min_child_samples=int(_params['min_child_samples']),
reg_alpha=_params['reg_alpha'],
reg_lambda=_params['reg_lambda'],
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
return XGBClassifier(
booster=_params['booster'],
eta=_params['eta'],
gamma=_params['gamma'],
max_depth=int(_params['max_depth']),
min_child_weight=int(_params['min_child_weight']),
max_delta_step=int(_params['max_delta_step']),
subsample=_params['subsample'],
colsample_bytree=_params['colsample_bytree'],
colsample_bylevel=_params['colsample_bylevel'],
colsample_bynode=_params['colsample_bynode'],
reg_lambda=_params['reg_lambda'],
alpha=_params['alpha'],
verbosity=0,
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params, _x, _y, _x_eval):
estimator_list = []
for idx, job in enumerate(self.cantidate_job_list):
if idx == _params['max_estimator']:
break
estimator_list.append(job.model.getModel(job.best_params))
s_train, s_test = stacking(
estimator_list,
_x,
_y,
_x_eval,
regression=False,
metric=accuracy_score,
stratified=_params['stratified'],
shuffle=_params['shuffle'],
random_state=0,
n_jobs=definitions.getNumberOfCore(),
)
def getModel(self, _params):
if _params['max_features'] == definitions.JSON_NONE:
_params['max_features'] = None
if _params['class_weight'] == definitions.JSON_NONE:
_params['class_weight'] = None
return ExtraTrees(
n_estimators= int(_params['n_estimators']),
criterion= _params['criterion'],
# max_depth= int(_params['max_depth']),
# min_samples_split= _params['min_samples_split'],
# min_samples_leaf= _params['min_samples_leaf'],
min_weight_fraction_leaf= _params['min_weight_fraction_leaf'],
max_features= _params['max_features'],
# max_leaf_nodes= int(_params['max_leaf_nodes']),
# min_impurity_decrease= _params['min_impurity_decrease'],
# bootstrap= _params['bootstrap'],
# bootstrap=True,
# oob_score= bool(_params['oob_score']),
class_weight= _params['class_weight'],
# ccp_alpha= _params['ccp_alpha']
n_jobs= definitions.getNumberOfCore(),
)
def __getRemovedOutlierIsolationForestDf(self, _df, _column_drop=True): iForest = IsolationForest(contamination=OUTLIER_FRACTION, n_jobs=definitions.getNumberOfCore()) return self.__getRemoveOutlierDf(_df, iForest, self.column_name_iforest, _column_drop)
def getIsolationForest(_df):
clf = IsolationForest(
contamination=OUTLIER_FRACTION,
n_jobs=definitions.getNumberOfCore()
)
return clf.fit_predict(_df)
def getLocalFactor(_df):
clf = LocalOutlierFactor(
contamination=OUTLIER_FRACTION,
n_jobs=definitions.getNumberOfCore()
)
return clf.fit_predict(_df)
def __getRemovedOutlierLocalFactorDf(self, _df):
localFactor = LocalOutlierFactor(contamination=OUTLIER_FRACTION,
n_jobs=definitions.getNumberOfCore())
return self.__getRemoveOutlierDf(_df, localFactor)
def __getRemovedOutlierIsolationForestDf(self, _df):
iForest = IsolationForest(contamination=OUTLIER_FRACTION,
n_jobs=definitions.getNumberOfCore())
return self.__getRemoveOutlierDf(_df, iForest)
def __getAddDbscanDf(self, _df):
df = self.__getDroppedColumnDf(_df.copy())
dbscan = DBSCAN(n_jobs=definitions.getNumberOfCore())
pred = dbscan.fit_predict(df)
df['added_cluster'] = pred
return df
def __getRemovedOutlierLocalFactorDf(self, _df, _column_drop=True): localFactor = LocalOutlierFactor(contamination=OUTLIER_FRACTION, n_jobs=definitions.getNumberOfCore()) return self.__getRemoveOutlierDf(_df, localFactor, self.column_name_localfactor, _column_drop)
