def MELHOR_RESULTADO_MG():
X = dataSet[:, 0:4]
y = dataSet[:, 4]
param = {
'learning_rate': [0.05],
'max_depth': [50],
'max_features': ['log2'],
'min_samples_leaf': [11],
'n_estimators': [60]
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=9)
model = GradientBoostingRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :", grid.best_score_)
print(grid.best_params_)
pltResults(best_model, 1, X_train, X_test, y_train, y_test)
#Conluido - Best Score: 0.73465 Semente: 9
#Best Params {'learning_rate': 0.05, 'max_depth': 50, 'max_features': 'log2', 'min_samples_leaf': 11, 'n_estimators': 60, 'random_state': 0}
#Best Score 0.7346493872397788
#Best Seed 9
#R2 Test: 0.7641468991808511 MSE Test: 0.0037531633086605658
#R2 Train: 0.8748115698686528 MSE Train: 0.003076693813010081
MELHOR_RESULTADO_MG()
def MELHOR_RESULTADO_NA():
X = dataSet[:, 0:4]
y = dataSet[:, 5]
param = {
'learning_rate': [0.05],
'max_depth': [20],
'max_features': ['log2'],
'min_samples_leaf': [5],
'n_estimators': [50],
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=1)
model = GradientBoostingRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :", grid.best_score_)
pltResults(best_model, 1, X_train, X_test, y_train, y_test)
#Conluido - Best Score: 0.73448 Semente: 9
#Best Params {'learning_rate': 0.05, 'max_depth': 10, 'max_features': 2, 'min_samples_leaf': 5, 'n_estimators': 50, 'random_state': 0}
#Best Score 0.734480868500437
MELHOR_RESULTADO_NA()
def MELHOR_RESULTADO_K():
X = dataSet[:, 0:4]
y = dataSet[:, 6]
param = {
'bootstrap': [True],
'max_depth': [10],
'max_features': ['log2'],
'min_samples_leaf': [2],
'n_estimators': [60]
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=3)
model = RandomForestRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :", grid.best_score_)
print(grid.best_params_)
pltResults(best_model, 0, X_train, X_test, y_train, y_test)
def MELHOR_RESULTADO_NA():
X = dataSet[:, 0:4]
y = dataSet[:, 5]
param = {
'bootstrap': [True],
'max_depth': [10],
'max_features': ['log2'],
'min_samples_leaf': [2],
'n_estimators': [100]
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=2)
model = RandomForestRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :", grid.best_score_)
print(grid.best_params_)
pltResults(best_model, 0, X_train, X_test, y_train, y_test)
#Conluido - Best Score: 0.61644 Semente: 2
#Best Params {'bootstrap': True, 'max_depth': 10, 'max_features': 'log2', 'min_samples_leaf': 2, 'n_estimators': 100}
#Best Score 0.6164423932855161
#Best Seed 2
#R2 Test: 0.6470743582563896 MSE Test: 0.011315230193482718
#R2 Train: 0.8844389141031836 MSE Train: 0.002520229138039719
MELHOR_RESULTADO_NA()
def MELHOR_RESULTADO_MG():
X = dataSet[:, 0:4]
y = dataSet[:, 4]
param = {
'bootstrap': [True],
'max_depth': [20],
'max_features': ['auto'],
'min_samples_leaf': [4],
'n_estimators': [60]
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=9)
model = RandomForestRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :", grid.best_score_)
print(grid.best_params_)
pltResults(best_model, 3, X_train, X_test, y_train, y_test)
#Conluido - Best Score: 0.73366 Semente: 9
#Best Params {'bootstrap': True, 'max_depth': 20, 'max_features': 'auto', 'min_samples_leaf': 4, 'n_estimators': 60}
#Best Score 0.7336590449066758
#Best Seed 9
#R2 Test: 0.7158820898963456 MSE Test: 0.00452120795457348
#R2 Train: 0.8717629515241512 MSE Train: 0.0031516181905177796
MELHOR_RESULTADO_MG()
def MELHOR_RESULTADO_K():
X = dataSet[:,0:4]
y = dataSet[:,6]
param ={'algorithm': ['brute'], 'n_neighbors': [8], 'weights': ['distance']}
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=4)
model = KNeighborsRegressor()
grid = GridSearchCV(model,param, cv=10,verbose=0,n_jobs=-1,scoring='r2',iid=True)
grid.fit(X_train,y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :",grid.best_score_)
print(grid.best_params_)
pltResults(best_model,3,X_train,X_test,y_train,y_test)
#Conluido - Best Score: 0.67596 Semente: 4
#Best Params {'algorithm': 'brute', 'n_neighbors': 8, 'weights': 'distance'}
#Best Score 0.6759638656720487
#Best Seed 4
#R2 Test: 0.8262489727826184 MSE Test: 0.0030635365836426814
#R2 Train: 0.9999999999999659 MSE Train: 8.640127461660114e-16
MELHOR_RESULTADO_K()
def MELHOR_RESULTADO_NA():
X = dataSet[:,0:4]
y = dataSet[:,5]
param ={'algorithm': ['brute'], 'n_neighbors': [6], 'weights': ['distance']}
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=9)
model = KNeighborsRegressor()
grid = GridSearchCV(model,param, cv=10,verbose=0,n_jobs=-1,scoring='r2',iid=True)
grid.fit(X_train,y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :",grid.best_score_)
print(grid.best_params_)
pltResults(best_model,3,X_train,X_test,y_train,y_test)
#Conluido - Best Score: 0.62096 Semente: 9
#Best Params {'algorithm': 'brute', 'n_neighbors': 6, 'weights': 'distance'}
#Best Score 0.6209552829789358
#Best Seed 9
#R2 Test: 0.7066316533521843 MSE Test: 0.009787010608152074
#R2 Train: 0.9999999999999105 MSE Train: 1.9219260943726335e-15
MELHOR_RESULTADO_NA()
def MELHOR_RESULTADO_MG():
X = dataSet[:,0:4]
y = dataSet[:,4]
param ={'algorithm': ['auto'], 'n_neighbors': [6], 'weights': ['distance']}
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=5)
model = KNeighborsRegressor()
grid = GridSearchCV(model,param, cv=10,verbose=0,n_jobs=-1,scoring='r2',iid=True)
grid.fit(X_train,y_train)
best_model = grid.best_estimator_
print("Cross Validation R2 Score :",grid.best_score_)
print(grid.best_params_)
pltResults(best_model,3,X_train,X_test,y_train,y_test)
#Conluido - Best Score: 0.69719 Semente: 4
#Best Params {'algorithm': 'auto', 'n_neighbors': 6, 'weights': 'distance'}
#Best Score 0.6971875382557432
#Best Seed 4
#R2 Test: 0.8477948730251221 MSE Test: 0.0026071842742281586
#R2 Train: 1.0 MSE Train: 0.0
MELHOR_RESULTADO_MG()
#dataSet = loadMainDataSet()
#Set features and target
y_column = 2
X = dataSet[:,0:2]
y = dataSet[:,2]
best_model , best_params, best_score,best_seed = findBalancedDataSet(range(1,10),X,y,GridSearchCVKNeighborsRegressor)
print("#Best Params",best_params)
print("#Best Score",best_score)
print("#Best Seed",best_seed)
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2,random_state=best_seed)
pltResults(best_model,X.shape[1]-1,X_train,X_test,y_train,y_test)
########## MG #################
#Conluido - Best Score: 0.69719 Semente: 4
#Best Params {'algorithm': 'auto', 'n_neighbors': 6, 'weights': 'distance'}
#Best Score 0.6971875382557432
#Best Seed 4
#R2 Test: 0.8477948730251221 MSE Test: 0.0026071842742281586
#R2 Train: 1.0 MSE Train: 0.0
def MELHOR_RESULTADO_MG():
X = dataSet[:,0:4]
y = dataSet[:,4]
param ={'algorithm': ['auto'], 'n_neighbors': [6], 'weights': ['distance']}
def MELHOR_RESULTADO_K():
X = dataSet[:, 0:4]
y = dataSet[:, 6]
param = {
'learning_rate': [0.05],
'max_depth': [30],
'max_features': ['log2'],
'min_samples_leaf': [4],
'n_estimators': [90],
}
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=3)
model = GradientBoostingRegressor(random_state=0)
grid = GridSearchCV(model,
param,
cv=10,
verbose=0,
n_jobs=-1,
scoring='r2',
iid=True)
grid.fit(X_train, y_train)
best_model = grid.best_estimator_
print(grid.best_score_)
print(grid.best_params_)
pltResults(best_model, 0, X_train, X_test, y_train, y_test)
#Conluido - Best Score: 0.63825 Semente: 3
#Best Params {'learning_rate': 0.05, 'max_depth': 30, 'max_features': 'log2', 'min_samples_leaf': 4, 'n_estimators': 90, 'random_state': 0}
#Best Score 0.6382454854124988
#R2 Test: 0.6772838242726009 MSE Test: 0.0078083038481006
#R2 Train: 0.9892885578561049 MSE Train: 0.00025464662511601085
MELHOR_RESULTADO_K()
############### MG ##########################
#Conluido - Best Score: 0.65662 Semente: 9
#Best Params {'learning_rate': 0.05, 'max_depth': 7, 'max_features': 'auto', 'min_samples_leaf': 5, 'n_estimators': 19, 'random_state': 0}
#Best Score 0.6566162097362611
#Best Seed 9
#R2 Test: 0.6698401069997604 MSE Test: 0.005253880453960908
#R2 Train: 0.7738415101819238 MSE Train: 0.0055581847751659365
#Conluido - Best Score: 0.73203 Semente: 9
#Best Params {'learning_rate': 0.05, 'max_depth': 50, 'max_features': 'log2', 'min_samples_leaf': 5, 'n_estimators': 50, 'random_state': 0}
#Best Score 0.7320304172385002
#Best Seed 9
#R2 Test: 0.6928402471363506 MSE Test: 0.004887876014100292
#R2 Train: 0.9412048892640388 MSE Train: 0.0014449782080243406
#Conluido - Best Score: 0.73465 Semente: 9
#Best Params {'learning_rate': 0.05, 'max_depth': 50, 'max_features': 'log2', 'min_samples_leaf': 11, 'n_estimators': 60, 'random_state': 0}
#Best Score 0.7346493872397788
#Best Seed 9
#R2 Test: 0.7641468991808511 MSE Test: 0.0037531633086605658
#R2 Train: 0.8748115698686528 MSE Train: 0.003076693813010081
############### Na ##########################
#Conluido - Best Score: 0.59157 Semente: 2
#Best Params {'learning_rate': 0.05, 'max_depth': 50, 'max_features': 'log2', 'min_samples_leaf': 5, 'n_estimators': 50, 'random_state': 0}
#Best Score 0.5915734707430875
#Best Seed 2
#R2 Test: 0.6002369455399681 MSE Test: 0.012816895257928434
#R2 Train: 0.9122383051241635 MSE Train: 0.0019139624633444869
#Conluido - Best Score: 0.59157 Semente: 2
#Best Params {'learning_rate': 0.05, 'max_depth': 20, 'max_features': 'log2', 'min_samples_leaf': 5, 'n_estimators': 50, 'random_state': 0}
#Best Score 0.5915734707430875
#Best Seed 2
#R2 Test: 0.6002369455399681 MSE Test: 0.012816895257928434
#R2 Train: 0.9122383051241635 MSE Train: 0.0019139624633444869
#Conluido - Best Score: 0.59717 Semente: 2
#Best Params {'learning_rate': 0.05, 'max_depth': 5, 'max_features': 'log2', 'min_samples_leaf': 5, 'n_estimators': 50, 'random_state': 0}
#Best Score 0.5971743288694583
#Best Seed 2
#R2 Test: 0.6153244736541088 MSE Test: 0.012333170548047006
#R2 Train: 0.8709102998267071 MSE Train: 0.002815269701498251
############### K #############################
#Conluido - Best Score: 0.63825 Semente: 3
#Best Params {'learning_rate': 0.05, 'max_depth': 30, 'max_features': 'log2', 'min_samples_leaf': 4, 'n_estimators': 90, 'random_state': 0}
#Best Score 0.6382454854124988
#Best Seed 3
#R2 Test: 0.6772838242726009 MSE Test: 0.0078083038481006
#R2 Train: 0.9892885578561049 MSE Train: 0.00025464662511601085