def linear():
params = get_params(request.args)
X_train, X_test, y_train, y_test = regressionData(params[2], params[3])
start, end = params[0], params[1]
model = linear_model.LinearRegression()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
return res.to_json(orient='index')
def ridge():
params = get_params(request.args)
start, end = params[0], params[1]
X_train, X_test, y_train, y_test = regressionData(params[2], params[3])
ridgereg = linear_model.Ridge(alpha=params[10], normalize=True)
ridgereg.fit(X_train, y_train)
y_pred = ridgereg.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("ridge")
print(params)
return res.to_json(orient='index')
def svm():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
clf = SVC(kernel=params[6])
clf.fit(X_train, np.ravel(y_train))
y_pred = clf.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("svm")
print(params)
return res.to_json(orient='index')
def naive():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
model = GaussianNB()
model.fit(X_train,np.ravel(y_train))
y_pred = model.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("naive")
print(params)
return res.to_json(orient='index')
def logistic():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
model = linear_model.LogisticRegression(random_state=0)
model.fit(X_train, np.ravel(y_train))
y_pred = model.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("log")
print(params)
return res.to_json(orient='index')
def dtree():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
clf = DecisionTreeClassifier(max_depth= params[7])
clf = clf.fit(X_train,np.ravel(y_train))
y_pred = clf.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("dtree")
print(params)
return res.to_json(orient='index')
def rtree():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
clf=RandomForestClassifier(n_estimators=params[8])
clf.fit(X_train,np.ravel(y_train))
y_pred=clf.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("rtree")
print(params)
return res.to_json(orient='index')
def knear():
params = get_params(request.args)
X_train, X_test, y_train, y_test = createData(params[2],params[3],params[4])
start, end = params[0], params[1]
classifier = KNeighborsClassifier(n_neighbors=params[5])
classifier.fit(X_train, np.ravel(y_train))
y_pred = classifier.predict(X_test)
res = result(X_test, y_test, y_pred)
res = res[start:end]
print("knear")
print(params)
return res.to_json(orient='index')
learning_rate=0.03, max_depth=5,
min_child_weight=80, n_estimators=1100,
reg_alpha=0.5, reg_lambda=0.7,
subsample=0.9, silent=True,
nthread=8, early_stopping_rounds=100)
'''
model_xgb = xgb.XGBRegressor(colsample_bytree=0.4603, gamma=0.0468,
learning_rate=0.05, max_depth=3,
min_child_weight=1.7817, n_estimators=2200,
reg_alpha=0.4640, reg_lambda=0.8571,
subsample=0.5213, silent=True,
nthread = 8)'''
print('\n\nStart...')
t0 = time.time()
train_preds, test_preds, mses = np.zeros(train.shape[0]), np.zeros((test.shape[0], 5)), []
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(n_splits=5, shuffle=True, random_state=520)
for i, (train_index, test_index) in enumerate(kf.split(train)):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
gbm = model_xgb.fit(train_feat1[predictors], train_feat1['血糖'])
predict = gbm.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(.5 * mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = gbm.predict(test[predictors])
cv = mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds)) * 0.5
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('XGBoost', mses, cv, test_preds)
import time
import numpy as np
from sklearn.model_selection import KFold
from sklearn.metrics import mean_squared_error
import function
train, test = function.read_file(path='a')
train["血糖"] = np.log1p(train["血糖"])
train, test = function.add_column(train, test)
train, test = function.transform(train, test)
print('\n\nStart...')
t0, mses = time.time(), []
train_preds, test_preds = np.zeros(train.shape[0]), np.zeros((test.shape[0], 5))
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(n_splits=5, shuffle=True, random_state=520)
for i, (train_index, test_index) in enumerate(kf.split(train)):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
gbm = function.settings.model_lgb.fit(train_feat1[predictors], train_feat1['血糖'],
categorical_feature=['性别', '体检日期'])
predict = gbm.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(.5 * mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = gbm.predict(test[predictors])
cv = .5 * mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds))
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('LGBM', mses, cv, test_preds)
train, test = function.read_file()
train["血糖"] = np.log1p(train["血糖"])
train, test = function.transform(train, test, fill_na=True)
pred = test.copy()
pred.drop('血糖', axis=1, inplace=True)
KRR = KernelRidge(kernel='polynomial', coef0=0, alpha=0.1, degree=2.1)
print('\n\nStart...')
t0 = time.time()
train_preds, test_preds, mses = np.zeros(train.shape[0]), np.zeros(
(test.shape[0], 5)), []
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(n_splits=5, shuffle=True, random_state=520)
for i, (train_index, test_index) in enumerate(kf.split(train)):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
gbm = KRR.fit(train_feat1[predictors], train_feat1['血糖'])
predict = gbm.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(
.5 *
mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = gbm.predict(test[predictors])
cv = mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds)) * 0.5
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('KRR', mses, cv, test_preds)
import time
import numpy as np
from sklearn.model_selection import KFold
from sklearn.metrics import mean_squared_error
import function
train, test = function.read_file()
train["血糖"] = np.log1p(train["血糖"])
train, test = function.transform(train, test)
print('\n\nStart...')
t0 = time.time()
train_preds, test_preds, mses = np.zeros(train.shape[0]), np.zeros((test.shape[0], 5)), []
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(n_splits=5, shuffle=True, random_state=520)
for i, (train_index, test_index) in enumerate(kf.split(train)):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
[train_feat1, train_feat2] = function.fill_data([train_feat1, train_feat2])
gbm = function.settings.model_xgb.fit(train_feat1[predictors], train_feat1['血糖'])
predict = gbm.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(.5 * mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = gbm.predict(test[predictors])
cv = mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds)) * 0.5
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('XGB_f', mses, cv, test_preds)
])
return self.meta_model_.predict(meta_features)
stacked_averaged_models = StackingAveragedModels(base_models=(ENet, GBoost,
KRR),
meta_model=lasso)
print('\n\nStart...')
t0 = time.time()
train_preds, test_preds, mses = np.zeros(train.shape[0]), np.zeros(
(test.shape[0], 5)), []
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(len(train), n_folds=5, shuffle=True, random_state=520)
for i, (train_index, test_index) in enumerate(kf):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
gbm = stacked_averaged_models.fit(train_feat1[predictors].values,
train_feat1['血糖'].values)
predict = gbm.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(
.5 *
mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = gbm.predict(test[predictors])
cv = mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds)) * 0.5
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('Stack', mses, cv, test_preds)
import function as f f.febre() f.mancha() f.dor_musc() f.dor_art() f.intens_art() f.edema_articulacao() f.conjuntivite() f.dor_cabeca() f.coceira() f.hiper_gang() f.disc_hemorr() f.neuro() f.result()
predictors = [f for f in test.columns if f not in ['血糖']]
kf = KFold(n_splits=5, shuffle=True, random_state=520)
# Instantiate a RandomRegressor object
MAXDEPTH = 40
regr = RandomForestRegressor(n_estimators=1000, # No of trees in forest
criterion="mse", # Can also be mae
max_features="sqrt", # no of features to consider for the best split
max_depth=MAXDEPTH, # maximum depth of the tree
min_samples_split=2, # minimum number of samples required to split an internal node
min_impurity_decrease=0, # Split node if impurity decreases greater than this value.
oob_score=True, # whether to use out-of-bag samples to estimate error on unseen data.
n_jobs=-1, # No of jobs to run in parallel
random_state=0,
verbose=0 # Controls verbosity of process
)
for i, (train_index, test_index) in enumerate(kf.split(train)):
print(' .{}/5.'.format(i + 1))
train_feat1, train_feat2 = train.iloc[train_index], train.iloc[test_index]
regr.fit(train_feat1[predictors], train_feat1['血糖'])
predict = predictions = regr.predict(train_feat2[predictors])
train_preds[test_index] += predict
mses.append(.5 * mean_squared_error(np.expm1(train_feat2['血糖']), np.expm1(predict)))
test_preds[:, i] = regr.predict(test[predictors])
cv = .5 * mean_squared_error(np.expm1(train['血糖']), np.expm1(train_preds))
print('\nFinished.\n\nSeconds -> %s\n' % str(time.time() - t0)[:8])
function.result('R.F.', mses, cv, test_preds)
