def parse(self, jsonData):
if jsonData['estimator'].endswith('Classifier'):
from sklearn.tree import DecisionTreeClassifier
self.estimator = DecisionTreeClassifier(criterion='gini')
self.is_regr = False
else:
from sklearn.tree import DecisionTreeRegressor
self.estimator = DecisionTreeRegressor(criterion='mse')
self.is_regr = True
self.params = {}
import sys
sys.path.insert(1, 'utils')
import dict_utils
self.params['max_depth'] = dict_utils.parse_related_properties(
'max_depth', jsonData, None)
self.params['min_samples_split'] = dict_utils.parse_related_properties(
'min_samples_split', jsonData, 2)
self.params['min_samples_leaf'] = dict_utils.parse_related_properties(
'min_samples_leaf', jsonData, 1)
self.params['max_leaf_nodes'] = dict_utils.parse_related_properties(
'max_leaf_nodes', jsonData, None)
sys.path.insert(1, 'output')
import DecisionTree_OM
self.output_manager = DecisionTree_OM.DecisionTree_OM(self)
def parse(self, jsonData):
if jsonData['estimator'] == 'LinearSVC':
from sklearn.svm import LinearSVC
self.estimator = LinearSVC(random_state=0)
self.is_regr = False
elif jsonData['estimator'] == 'LinearSVR':
from sklearn.svm import SVR
self.estimator = SVR(kernel='linear')
self.is_regr = True
self.params = {}
self.params['C'] = dict_utils.parse_related_properties('C_exp',
jsonData,
0,
is_exp=True)
'''
if 'C' in jsonData:
self.params['C'] = [jsonData['C']]
elif 'C_lower_exp' in jsonData and 'C_upper_exp' in jsonData:
import numpy as np
l = jsonData['C_lower_exp']
u = jsonData['C_upper_exp']
self.params['C'] = np.logspace(l, u, u-l+1)
elif 'C_lower_exp' in jsonData:
self.params['C'] = [jsonData['C_lower_exp']]
elif 'C_upper_exp' in jsonData:
self.params['C'] = [jsonData['C_upper_exp']]
else:
self.params['C'] = 1
'''
import sys
sys.path.insert(1, 'output')
import SVM_OM
self.output_manager = SVM_OM.SVM_OM(self)
def parse(self, jsonData):
if jsonData['estimator'].endswith('Classifier'):
from sklearn.ensemble import RandomForestClassifier
self.estimator = RandomForestClassifier(criterion='gini')
self.is_regr = False
else:
from sklearn.ensemble import RandomForestRegressor
self.estimator = RandomForestRegressor(criterion='mse')
self.is_regr = True
self.params = {}
import sys
sys.path.insert(1, 'utils')
import dict_utils
self.params['n_estimators'] = dict_utils.parse_related_properties(
'n_estimators', jsonData, 100)
self.params['max_depth'] = dict_utils.parse_related_properties(
'max_depth', jsonData, None)
self.params['min_samples_split'] = dict_utils.parse_related_properties(
'min_samples_split', jsonData, 2)
self.params['min_samples_leaf'] = dict_utils.parse_related_properties(
'min_samples_leaf', jsonData, 1)
self.params[
'min_weight_fraction_leaf'] = dict_utils.parse_related_properties(
'min_weight_fraction_leaf', jsonData, 0.0)
self.params['max_features'] = dict_utils.parse_related_properties(
'max_features', jsonData, 'auto')
self.params['max_leaf_nodes'] = dict_utils.parse_related_properties(
'max_leaf_nodes', jsonData, None)
self.params[
'min_impurity_decrease'] = dict_utils.parse_related_properties(
'min_impurity_decrease', jsonData, 0.0)
self.params['n_jobs'] = dict_utils.parse_related_properties(
'n_jobs', jsonData, None)
self.params['verbose'] = dict_utils.parse_related_properties(
'verbose', jsonData, 0)
sys.path.insert(1, 'output')
import RandomForest_OM
self.output_manager = RandomForest_OM.RandomForest_OM(self)
def parse(self, jsonData):
if jsonData['estimator'].endswith('Classifier'):
from sklearn.neighbors import KNeighborsClassifier
self.estimator = KNeighborsClassifier()
self.is_regr = False
else:
from sklearn.neighbors import KNeighborsRegressor
self.estimator = KNeighborsRegressor()
self.is_regr = True
self.params = {}
import sys
sys.path.insert(1, 'utils')
import dict_utils
self.params['n_neighbors'] = dict_utils.parse_related_properties(
'n_neighbors', jsonData, 5)
sys.path.insert(1, 'output')
import Knn_OM #as Knn_OM
self.output_manager = Knn_OM.Knn_OM(self)