class XGB(BaseModel):
def __init__(self):
self.clf = XGBClassifier(
n_estimators=200,
max_depth=20,
learning_rate=0.1,
random_state=0,
booster="gbtree",
use_label_encoder=False,
)
def train(self, X_train, Y_train):
X_train, Y_train = do_rebalance(X_train, Y_train)
self.clf.fit(X_train, Y_train)
def test(self, X_test, Y_test):
Y_prob = self.clf.predict_proba(X_test)
auc = metrics.roc_auc_score(Y_test, Y_prob[:, 1])
def predict(self, X):
Y_prob = self.clf.predict_proba(X)
return Y_prob
def load_model(self, model_path):
self.clf.load_model(model_path)
# with open(model_path, "rb+") as file:
# self.clf = pickle.load(file)
def save_model(self, model_path):
self.clf.save_model(model_path)
def train(file):
y_data=makeRawDataset(file) #provides with raw data
X=pd.read_csv(file).Question
X_without=removePunc(X) #without punctuations
X_correct=fuzzy(X_without) #with fuzzy
X_enc=encode(X_correct)
labels=[">","<","<=",">=","==","NULL","LIKE"]
encoder=LabelEncoder()
codes=encoder.fit_transform(labels)
codeMap={labels[i]:codes[i] for i in range(len(labels))}
inverseMap={codes[i]:labels[i] for i in range(len(labels))}
maps={"codeMap":codeMap,"inverseMap":inverseMap}
np.save("Map.npy",maps)
y=[]
for i in y_data:
y.append(codeMap[i])
X_train, X_test, y_train, y_test = train_test_split(X_enc, y, test_size = 0.2, random_state = 42)
model = XGBClassifier()
model.fit(X_train,y_train)
model.save_model("WhereCond.model")
y_hat=model.predict(X_test)
print(y_hat[:10])
y_pred=[]
for i in y_hat:
y_pred.append(inverseMap[i])
y_true=[]
for i in y_test:
y_true.append(inverseMap[i])
sk_report = classification_report(digits=6,y_true=y_test,y_pred=y_hat)
print(sk_report)
def tree_optimization(learning_rate, gamma, max_depth, subsample, reg_lambda,
num_parallel_tree, min_child_weight):
global tree_no, best_score
X_train, X_eval, y_train, y_eval = train_test_split(features,
labels,
test_size=0.2,
shuffle=True,
stratify=labels)
X_train, X_test, y_train, y_test = train_test_split(X_train,
y_train,
test_size=0.2,
shuffle=True,
stratify=y_train)
booster_params = {
'n_estimators': 500,
'learning_rate': learning_rate,
'gamma': gamma,
'max_depth': int(np.around(max_depth)),
'subsample': subsample,
'sampling_method': 'gradient_based',
'reg_lambda': reg_lambda,
'min_child_weight': int(np.around(min_child_weight)),
'num_parallel_tree': int(np.around(num_parallel_tree)),
'objective': 'binary:logistic',
'verbosity': 1,
'max_delta_step ': 1
}
print("generating model")
trip_model = XGBClassifier(**booster_params)
trip_model.fit(X=X_train,
y=y_train,
eval_set=[(X_test, y_test)],
eval_metric=f1_eval,
early_stopping_rounds=25)
print("Training Accuracy: %.2f" % (trip_model.score(X_eval, y_eval) * 100),
"%")
preds = trip_model.predict(X_eval)
current_score = f1_score(y_eval, preds)
if current_score > best_score:
print(f"Score increased to {current_score} from {best_score}")
best_score = current_score
sub_pred = trip_model.predict(sub)
trip_model.save_model(
f'logs/f1_{run_num}_{best_score}_{tree_no}_thresh.model')
savetxt(f'logs/{run_num}_{tree_no}_preds.txt', sub_pred, delimiter=',')
booster_params['f1'] = current_score
booster_params['tree_no'] = tree_no
record_history(booster_params)
tree_no += 1
return current_score
def xgboost(self, x_train, y_train, x_valid, y_valid):
xgb_model = XGBClassifier(objective='multi:softmax',
verbose=2)
eval_set = [(x_valid, y_valid)]
xgb_model.fit(x_train, y_train, eval_set=eval_set, eval_metric='merror', verbose=True, early_stopping_rounds=10)
plot_importance(xgb_model)
pyplot.show()
xgb_model.save_model('./modfile/tf_idf_XGBoost.model')
return xgb_model, x_valid, y_valid
def create_tree():
X_train, X_eval, y_train, y_eval = train_test_split(
features,
labels,
test_size=0.2,
shuffle=True,
stratify=labels,
random_state=RANDOM_SEED)
X_train, X_test, y_train, y_test = train_test_split(
X_train,
y_train,
test_size=0.2,
shuffle=True,
stratify=y_train,
random_state=RANDOM_SEED)
booster_params = {
'n_estimators': 500,
'objective': 'binary:logistic',
'verbosity': 1
}
print("generating model")
trip_model = XGBClassifier(**booster_params)
trip_model.fit(X=X_train,
y=y_train,
eval_set=[(X_test, y_test)],
eval_metric=f1_eval,
early_stopping_rounds=50)
iter_num = trip_model.best_iteration + 30
booster_params = {
'n_estimators': iter_num,
'objective': 'binary:logistic',
'verbosity': 1
}
trip_model = XGBClassifier(**booster_params)
trip_model.fit(X=X_train,
y=y_train,
eval_set=[(X_test, y_test)],
eval_metric=f1_eval)
print("Training Accuracy: %.2f" % (trip_model.score(X_eval, y_eval) * 100),
"%")
preds = trip_model.predict(X_eval)
current_score = f1_score(y_eval, preds)
trip_model.save_model(f'logs/f1_{current_score}_thresh.model')
sub_pred = trip_model.predict(sub)
savetxt(f'logs/{current_score}_preds.txt', sub_pred, delimiter=',')
return trip_model
def train_xgboost(self):
"""
Train an XGBoost classifier on the dataset and save model as artifact
"""
from xgboost import XGBClassifier
xgb = XGBClassifier()
xgb.fit(self.X, self.y)
xgb.save_model(script_path("model.bst"))
with open(script_path("model.bst"), "rb") as fh:
self.buffered_xgb_model = fh.read()
self.next(self.join)
def scikitAPI(trainFile):
X, Y, X_train, X_test, Y_train, Y_test = inputData(trainFile)
'''
Y = transLabel(Y)[1]
Y_train = transLabel(Y_train)[1]
Y_test = transLabel(Y_test)[1]
'''
start = time.time()
print("class: {0}".format(np.unique(Y_train)))
print("train data shape: %r, train target shape: %r" %
(X_train.shape, Y_train.shape))
print("test data shape: %r, test target shape: %r" %
(X_test.shape, Y_test.shape))
#params = {"n_estimators":300, "num_leaves":128, "learning_rate":0.1}
params = {"n_estimators": 50, "learning_rate": 0.1}
print("{:-<50}".format(""))
print("params", params)
#clf_test = LGBMClassifier(n_estimators=200)
# clf_test = XGBClassifier(**params)
# clf_test.fit(X_train, Y_train)
# print('Training time:{0}'.format(time.time()-start))
# print("clf class: ",clf_test.classes_)
# #pred = clf_test.predict(X_test)
# #print(accuracy_score(pred, y_test))
# print("Traing Acc: ", clf_test.score(X_train, Y_train), np.shape(X_train))
# print("Test Acc: ", clf_test.score(X_test, Y_test), np.shape(X_test))
# print("Total Acc: :", clf_test.score(X, Y), np.shape(X))
xgtrain = xgb.DMatrix(X, Y)
clf = XGBClassifier(**params).fit(X, Y)
model_path = 'xgb/model.joblib'
# joblib.dump(clf, model_path, compress=1)
#clf.save_model("lgbm_model.ml")
model_path = 'xgb/model.bst'
clf.save_model(model_path)
# clf.dump_model('xgb/dump.raw.txt', 'xgb/featmap.txt')
# print("ACC: ", load_model(model_path).score(X_test, Y_test))
print("model save to {}".format(model_path))
print("model.ml size: {:.3f} KB".format(
os.path.getsize(model_path) / 1024))
clf = load_model(model_path, Y)
print(clf.predict([np.arange(51)]))
def train_lazy():
# Load the dataset
X, y = load_data()
# Split the data
X_train, X_val, y_train, y_val = split_dataset(X, y)
# # Normalize
X_train = normalize(X_train)
X_val = normalize(X_val)
# uncomment to check the performance of the 25 models
# clf = LazyClassifier(verbose=0,ignore_warnings=True, custom_metric=None)
# # fit
# scores,_ = clf.fit(X_train, X_val, y_train, y_val)
# # print
# print(scores)
# Final model
# check if model exist
if os.path.isfile(config.MODEL_PATH):
model = XGBClassifier()
model.load_model(config.MODEL_PATH)
else:
model = XGBClassifier()
model.fit(X_train,
y_train,
eval_metric="error",
eval_set=[(X_train, y_train), (X_val, y_val)],
verbose=True)
# save model
model.save_model(config.MODEL_PATH)
# performance on train set
y_pred = model.predict(X_train)
# evaluate predictions
print_performance(y_train, y_pred, 'train')
# performance on val set
y_pred = model.predict(X_val)
# evaluate predictions
print_performance(y_val, y_pred, 'val')
# Load the test dataset
X_test, y_test = load_test_data()
# # Normalize
X_test = normalize(X_test)
# get prediction
y_pred = model.predict(X_test)
# evaluate predictions
print_performance(y_test, y_pred, 'test')
# print
plot_performance(model)
def build():
data = load_kickstarter(DATA_PATH)
encoder, data = preprocess(data)
x_train, x_test, y_train, y_test = train_test_split(data.drop(['state'], axis=1), data.state, test_size=0.2, random_state=SEED)
clf = XGBClassifier(n_jobs=6, random_state=SEED)
clf.fit(x_train, y_train, verbose=True)
dump(encoder, 'onehot-150k.joblib')
clf.save_model('xgb-150k-v1.model')
def train_model(self):
"""Trains the model
Returns
-------
testing and training f1 scores
"""
df = Crunchbase().format_crunchbase()
subset = df[['type_Group B', 'money_raised_at_ipo', 'number_of_acquisitions', 'valuation_at_ipo',
'employee_cat', 'number_of_lead_investors', 'number_of_lead_investments', 'number_of_investments',
'type_Group A', 'industries_type_0', 'number_of_employee_profiles',
'number_of_events',
'number_of_investors', 'total_products_active', 'type_For Profit', 'ipo_status']]
x = subset[subset.columns.drop('ipo_status')]
y = subset[['ipo_status']]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=2)
# scale data
scaler = StandardScaler()
# save the scaler
x_train_scaled = scaler.fit_transform(x_train)
pickle.dump(scaler, open('./Model/scaler.pkl', 'wb'))
x_test_scaled = scaler.transform(x_test)
# scale data
scaler = StandardScaler()
# save the scaler
x_train_scaled = scaler.fit_transform(x_train)
pickle.dump(scaler, open('./Model/scaler.pkl', 'wb'))
x_test_scaled = scaler.transform(x_test)
# transform data to balance it
smote = SMOTE(random_state=42, sampling_strategy=0.5)
x_train_scaled_smote, y_train_smote = smote.fit_resample(x_train_scaled, y_train)
xgb = XGBClassifier()
xgb.fit(x_train_scaled_smote, y_train_smote)
# save the trained model
xgb.save_model('./Model/final_model.json')
y_train_pred_xgb = xgb.predict(x_train_scaled_smote)
y_test_pred_xgb = xgb.predict(x_test_scaled)
training_f1_score = f1_score(y_train_smote, y_train_pred_xgb)
testing_f1_score = f1_score(y_test, y_test_pred_xgb)
return training_f1_score, testing_f1_score
def train_model():
x, y = load_data()
model = create_dl_model()
model.fit(x,
y,
epochs=20,
batch_size=32,
verbose=0,
callbacks=keras.callbacks.EarlyStopping(monitor='loss',
patience=3))
layermodel_1 = keras.models.Model(inputs=model.input,
outputs=model.layers[2].output)
x = layermodel_1.predict(x)
xgb = XGBClassifier(n_estimators=110, learning_rate=0.12)
xgb.fit(x, y.ravel())
layermodel_1.save("../files/DNN.h5")
xgb.save_model("../files/xgboost.model")
def train(X_train, y_train, X_test, y_test, i):
if i == 0:
name = 'citi'
elif i == 1:
name = 'jpm'
else:
name = 'boa'
weights = (y_train == 0).sum()/(1.0 * (y_train == 1).sum())
model = XGBClassifier(tree_method='gpu_hist', objective='binary:logistic',
n_estimators=300, scale_pos_weight=weights, n_jobs=6)
if name == 'citi':
model.fit(X_train, y_train, eval_metric=['auc'])
else:
model.fit(X_train, y_train, eval_metric=['auc'])
y_pred = model.predict(X_test)
get_results(y_pred, y_test)
model.save_model(f'{name}.model')
return model
def xgbmodel():
data1 = pd.read_csv('/home/msmal/object_type/X_enc_object.csv', header=0)
print("data1 read over!")
data2 = pd.read_csv('/home/msmal/float_type/train.csv', header=0)
data2[np.isinf(data2)] = -1
data2[np.isnan(data2)] = -2
print("data2 read over!")
X = pd.concat([data1.iloc[:, 1:], data2], axis=1)
y = pd.read_csv('/home/msmal/label.csv', header=0)
scaler = preprocessing.StandardScaler()
X_scaled = scaler.fit_transform(X)
encoder = LabelEncoder()
y = encoder.fit_transform(y)
X_train, X_test, Y_train, Y_test = train_test_split(X_scaled, y, test_size=0.005, random_state=49)
#xgbmodel:xgb = XGBClassifier(n_jobs=-1)
xgb = XGBClassifier(n_jobs=-1,n_estimators=500,learning_rate=0.05,subsample=0.8)
xgb.fit(X_train, Y_train)
xgb.save_model('xgb1.model')
print("model has saved!")
xgbpre = xgb.predict(X_test)
xgb_report = metrics.classification_report(Y_test, xgbpre)
print(xgb_report)
thresholds = np.sort(model.feature_importances_)
print(thresholds)
for thresh in thresholds:
selection = SelectFromModel(model, threshold=thresh, prefit = True)
parameter = {
'n_estimators': [100, 200, 400],
'learning_rate' : [0.03, 0.05, 0.07, 0.1],
'colsample_bytree': [0.6, 0.7, 0.8, 0.9],
'colsample_bylevel':[0.6, 0.7, 0.8, 0.9],
'max_depth': [4, 5, 6]
}
search = GridSearchCV( XGBRegressor(), parameter, cv =5, n_jobs = -1)
select_x_train = selection.transform(x_train)
search.fit(select_x_train, y_train)
select_x_test = selection.transform(x_test)
x_pred = search.predict(select_x_test)
score = r2_score(y_test, x_pred)
print('R2는',score)
print("Thresh=%.3f, n=%d, R2: %.2f%%" %(thresh, select_x_train.shape[1], score*100.0))
model.save_model('./model/xgb_save/m34sfm_cancer/cancer.xgb' + str(select_x_train.shape[1])+'.model')
'''
def train_xgboost(data: IrisData):
clf = XGBClassifier()
clf.fit(data.X, data.y)
model_path = os.path.join(XGBoostFolder, "model.json")
clf.save_model(model_path)
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
from sklearn.preprocessing import OneHotEncoder
from sklearn.compose import ColumnTransformer
def get_data():
url = 'simple_data.csv'
return pd.read_csv(url)
columnTransformer = ColumnTransformer([('encoder', OneHotEncoder(), [1])], remainder='passthrough')
df = get_data()
df['position'] = df['position'].str.extract(r'(\w+),?')
X = df.drop(columns=['all_star'])
X = columnTransformer.fit_transform(X)
y = df.all_star
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=99)
model = XGBClassifier(objective='binary:logistic', use_label_encoder=False)
model.fit(X_train, y_train)
model.save_model('xgb.model')
predictions = model.predict(X_test)
print("XGBoost Training Accuracy")
print(f'Accuracy: {round(accuracy_score(y_test, predictions) * 100, 3)}%')
print(classification_report(y_test, predictions))
x_train, x_test, y_train, y_test = train_test_split(x,
y,
train_size=0.8,
random_state=77,
shuffle=True)
# 2. 모델
model = XGBClassifier(n_estimators=1000, learning_rate=0.1)
# model = XGBRegressor(learning_rate=0.01)
# 3. 훈련
model.fit(
x_train,
y_train,
verbose=1, #다 보여 준다(0 / 1 , False / True)
eval_metric='logloss', #keras의 metrics와 동일
eval_set=[(x_train, y_train), (x_test, y_test)])
# 4. 평가, 예측
result = model.evals_result()
# print("eval's results : ", result)
acc = model.score(x_test, y_test)
print('acc:', acc)
y_pred = model.predict(x_test)
print('최종정답률:', accuracy_score(y_test, y_pred))
model.save_model('./save/xgb_save/cancer2.xgb.model')
print('저장완료')
selection_x_train = selection.transform(x_train)
selection_x_test = selection.transform(x_test)
print(selection_x_train.shape)
selection_model = XGBClassifier(n_estimators=1000,
max_depth=4,
learning_rate=0.5,
n_jobs=-1)
selection_model.fit(selection_x_train,
y_train,
verbose=False,
eval_metric=["merror", "mlogloss"],
eval_set=[(selection_x_train, y_train),
(selection_x_test, y_test)],
early_stopping_rounds=100)
y_pred = selection_model.predict(selection_x_test)
# results = selection_model.evals_result()
# print("evals_result : \n", results)
score = accuracy_score(y_test, y_pred)
print("Thresh=%.3f, n=%d, acc: %.2f%%" %
(thresh, selection_x_train.shape[1], score * 100.0))
# (120, 1)
# Thresh=0.621, n=1, acc: 96.67%
model.save_model("./model/xgb_save/iris_acc_96.67_model")
# 不可视化数据集loss
# model = XGBClassifier()
# model.fit(X_train, y_train)
##可视化测试集的loss
model = XGBClassifier()
eval_set = [(X_test, y_test)]
model.fit(X_train,
y_train,
early_stopping_rounds=100,
eval_metric="logloss",
eval_set=eval_set,
verbose=True)
# 改为True就能可视化loss
model.save_model("00001.model")
model.fit(X, Y)
plot_importance(model)
pyplot.show()
y_pred = model.predict(X_test)
predictions = [round(value) for value in y_pred]
accuracy = accuracy_score(y_test, predictions)
print("Accuracy: %.2f%%" % (accuracy * 100.0))
##Accuracy: 77.56%
test_auc2 = roc_auc_score(y_test, y_pred) #验证集上的auc值
print("xgb_muliclass_auc:", test_auc2)
X = df_2018.drop("job_title", axis=1)
y = df_2018["job_title"]
# Splitting data into training and test set
X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.80, test_size=0.20)
# save out the split training data to use with Cloud AutoML
with open("train_data_2018.csv", "+w") as file:
pd.concat([X_train, y_train], axis=1).to_csv(file, index=False)
with open("test_data_2018.csv", "+w") as file:
pd.concat([X_test, y_test], axis=1).to_csv(file, index=False)
# encode all features using ordinal encoding
encoder_x = ce.OrdinalEncoder()
X_encoded = encoder_x.fit_transform(X)
# you'll need to use a different encoder for each dataframe
encoder_y = ce.OrdinalEncoder()
y_encoded = encoder_y.fit_transform(y)
# split encoded dataset
X_train_encoded, X_test_encoded, y_train_encoded, y_test_encoded = train_test_split(X_encoded, y_encoded,
train_size=0.80, test_size=0.20)
from xgboost import XGBClassifier
# train XGBoost model with default parameters
my_model = XGBClassifier()
my_model.fit(X_train_encoded, y_train_encoded, verbose=False)
# and save our model
my_model.save_model("xgboost_baseline.model")
for thres in threshold:
selection = SelectFromModel(model, threshold=thres, prefit=True)
select_x_train = selection.transform(x_train)
select_x_test = selection.transform(x_test)
selection_model = LGBMClassifier(n_estimators=100,
learning_rate=0.05,
n_jobs=-1)
selection_model.fit(select_x_train,
y_train,
verbose=False,
eval_metric=['logloss', 'error'],
eval_set=[(select_x_train, y_train),
(select_x_test, y_test)],
early_stopping_rounds=20)
y_pred = selection_model.predict(select_x_test)
acc = accuracy_score(y_test, y_pred)
print("Thresh=%.3f, n = %d, ACC : %.2f%%" %
(thres, select_x_train.shape[1], acc * 100.0))
# result = selection_model.evals_result()
# print("eval's result : ", result)
model.save_model("./model/sample/cancer/cancer_rmse=%.3f-r2=%.2f.model" %
(thres, acc))
class XGBoost(BaseAlgorithm):
def __init__(self, algorithm_settings, problem_type):
super().__init__(algorithm_settings)
self.problem_type = problem_type
def build(self):
if self.problem_type == SupervisedTask.regression:
self.build_regression_model()
elif self.problem_type == SupervisedTask.classification:
self.build_classification_model()
else:
raise TypeError('Unknown problem_type')
def build_regression_model(self):
from xgboost import XGBRegressor
self.model = XGBRegressor(
max_depth=self.algorithm_settings.max_depth,
learning_rate=self.algorithm_settings.learning_rate,
n_estimators=self.algorithm_settings.n_estimators,
objective=self.algorithm_settings.objective,
booster=self.algorithm_settings.booster,
n_jobs=self.algorithm_settings.n_jobs,
gamma=self.algorithm_settings.gamma,
min_child_weight=self.algorithm_settings.min_child_weight,
max_delta_step=self.algorithm_settings.max_delta_step,
subsample=self.algorithm_settings.subsample,
reg_alpha=self.algorithm_settings.reg_alpha,
reg_lambda=self.algorithm_settings.reg_lambda,
random_state=self.algorithm_settings.random_state)
def build_classification_model(self):
from xgboost import XGBClassifier
self.model = XGBClassifier(
max_depth=self.algorithm_settings.max_depth,
learning_rate=self.algorithm_settings.learning_rate,
n_estimators=self.algorithm_settings.n_estimators,
objective=self.algorithm_settings.objective,
booster=self.algorithm_settings.booster,
n_jobs=self.algorithm_settings.n_jobs,
gamma=self.algorithm_settings.gamma,
min_child_weight=self.algorithm_settings.min_child_weight,
max_delta_step=self.algorithm_settings.max_delta_step,
subsample=self.algorithm_settings.subsample,
reg_alpha=self.algorithm_settings.reg_alpha,
reg_lambda=self.algorithm_settings.reg_lambda,
random_state=self.algorithm_settings.random_state)
def train(self, train_x, train_y, settings):
self.model.fit(train_x,
train_y,
eval_metric=self.algorithm_settings.eval_metric)
self.save(settings)
def evaluate(self, test_x):
prediction = self.model.predict(test_x)
prediction = prediction.reshape(-1, 1)
return prediction
def load(self, model_path):
self.model.load_model(fname=model_path)
def save(self, settings):
model_save_dir = os.path.join(settings.models_path, 'xgboost_models')
os.makedirs(model_save_dir, exist_ok=True)
model_name = self.get_model_name(settings)
save_path = os.path.join(model_save_dir, model_name)
self.model.save_model(fname=save_path)
print(f"Model saved to: {save_path}")
def get_model_name(self, settings):
if settings.problem_type == SupervisedTask.regression:
return 'regression_model.xgb'
else:
return 'classification_model.xgb'
def train_and_generate_model():
#global log_fd
global log_fd_opt
global tr_input_arr
global tr_angle_arr
global val_input_arr
global val_angle_arr
data_len = len(exchange_rates)
log_fd_tr = open("./train_progress_log_" +
dt.now().strftime("%Y-%m-%d_%H-%M-%S") + ".txt",
mode="w")
# inner logger function for backtest
def logfile_writeln_tr(log_str):
nonlocal log_fd_tr
log_fd_tr.write(log_str + "\n")
log_fd_tr.flush()
print("data size of rates: " + str(data_len))
print("num of rate datas for tarin: " +
str(COMPETITION_TRAIN_DATA_NUM_AT_RATE_ARR))
print("input features sets for tarin: " + str(COMPETITION_TRAIN_DATA_NUM))
logfile_writeln_tr("data size of rates: " + str(data_len))
logfile_writeln_tr("num of rate datas for tarin: " +
str(COMPETITION_TRAIN_DATA_NUM_AT_RATE_ARR))
tr_input_mat = []
tr_angle_mat = []
is_loaded_input_mat = False
if os.path.exists("./tr_input_mat.pickle"):
with open('./tr_input_mat.pickle', 'rb') as f:
tr_input_mat = pickle.load(f)
with open('./tr_angle_mat.pickle', 'rb') as f:
tr_angle_mat = pickle.load(f)
is_loaded_input_mat = True
else:
for i in range(DATA_HEAD_ASOBI,
len(exchange_rates) - DATA_HEAD_ASOBI - OUTPUT_LEN,
SLIDE_IDX_NUM_AT_GEN_INPUTS_AND_COLLECT_LABELS):
tr_input_mat.append([
exchange_rates[i],
(exchange_rates[i] - exchange_rates[i - 1]) /
exchange_rates[i - 1],
get_rsi(exchange_rates, i),
get_ma(exchange_rates, i),
get_ma_kairi(exchange_rates, i),
get_bb_1(exchange_rates, i),
get_bb_2(exchange_rates, i),
get_ema(exchange_rates, i),
get_ema_rsi(exchange_rates, i),
get_cci(exchange_rates, i),
get_mo(exchange_rates, i),
get_lw(exchange_rates, i),
get_ss(exchange_rates, i),
get_dmi(exchange_rates, i),
get_vorarity(exchange_rates, i),
get_macd(exchange_rates, i),
str(judge_chart_type(exchange_rates[i - CHART_TYPE_JDG_LEN:i]))
])
tr_input_mat.append([
reverse_exchange_rates[i],
(reverse_exchange_rates[i] - reverse_exchange_rates[i - 1]) /
reverse_exchange_rates[i - 1],
get_rsi(reverse_exchange_rates, i),
get_ma(reverse_exchange_rates, i),
get_ma_kairi(reverse_exchange_rates, i),
get_bb_1(reverse_exchange_rates, i),
get_bb_2(reverse_exchange_rates, i),
get_ema(reverse_exchange_rates, i),
get_ema_rsi(reverse_exchange_rates, i),
get_cci(reverse_exchange_rates, i),
get_mo(reverse_exchange_rates, i),
get_lw(reverse_exchange_rates, i),
get_ss(reverse_exchange_rates, i),
get_dmi(reverse_exchange_rates, i),
get_vorarity(reverse_exchange_rates, i),
get_macd(reverse_exchange_rates, i),
str(
judge_chart_type(
reverse_exchange_rates[i - CHART_TYPE_JDG_LEN:i]))
])
tmp = exchange_rates[i + OUTPUT_LEN] - exchange_rates[i]
if tmp >= 0:
tr_angle_mat.append(1)
else:
tr_angle_mat.append(0)
tmp = reverse_exchange_rates[
i + OUTPUT_LEN] - reverse_exchange_rates[i]
if tmp >= 0:
tr_angle_mat.append(1)
else:
tr_angle_mat.append(0)
if is_loaded_input_mat == False:
with open('tr_input_mat.pickle', 'wb') as f:
pickle.dump(tr_input_mat, f)
with open('tr_angle_mat.pickle', 'wb') as f:
pickle.dump(tr_angle_mat, f)
#log output for tensorboard
#configure("logs/xgboost_trade_cpu_1")
tr_input_arr = np.array(tr_input_mat[0:COMPETITION_TRAIN_DATA_NUM])
tr_angle_arr = np.array(tr_angle_mat[0:COMPETITION_TRAIN_DATA_NUM])
watchlist = None
split_idx = COMPETITION_TRAIN_DATA_NUM + int(
(len(tr_input_mat) - COMPETITION_TRAIN_DATA_NUM) *
VALIDATION_DATA_RATIO)
if VALIDATION_DATA_RATIO != 0.0:
val_input_arr = np.array(
tr_input_mat[COMPETITION_TRAIN_DATA_NUM:split_idx])
val_angle_arr = np.array(
tr_angle_mat[COMPETITION_TRAIN_DATA_NUM:split_idx])
watchlist = [(tr_input_arr, tr_angle_arr),
(val_input_arr, val_angle_arr)]
else:
watchlist = [(tr_input_arr, tr_angle_arr)]
start = time.time()
if is_param_tune_with_optuna:
log_fd_opt = open("./tune_progress_log_" +
dt.now().strftime("%Y-%m-%d_%H-%M-%S") + ".txt",
mode="w")
study = None
if is_use_db_at_tune:
study = optuna.Study(study_name='fxsystrade',
storage='sqlite:///../fxsystrade.db')
else:
study = optuna.create_study()
parallel_num = RAPTOP_THREAD_NUM * 2
if is_colab_cpu or is_exec_at_mba:
parallel_num = COLAB_CPU_AND_MBA_THREAD_NUM * 2
if special_optuna_parallel_num != -1:
parallel_num = special_optuna_parallel_num
study.optimize(opt, n_trials=OPTUNA_TRIAL_NUM, n_jobs=parallel_num)
process_time = time.time() - start
logfile_writeln_opt("best_params: " + str(study.best_params))
logfile_writeln_opt("best_value: " + str(study.best_value))
logfile_writeln_opt("best_trial: " + str(study.best_trial))
logfile_writeln_opt("excecution time of tune: " + str(process_time))
log_fd_opt.flush()
log_fd_opt.close()
exit()
param = {}
n_thread = RAPTOP_THREAD_NUM
if is_use_gpu:
param['tree_method'] = 'gpu_hist'
param['max_bin'] = 16
param['gpu_id'] = 0
n_thread = COLAB_CPU_AND_MBA_THREAD_NUM
if is_colab_cpu or is_exec_at_mba:
n_thread = COLAB_CPU_AND_MBA_THREAD_NUM
logfile_writeln_tr("training parameters are below...")
logfile_writeln_tr(str(param))
eval_result_dic = {}
logfile_writeln_tr("num_round: " + str(NUM_ROUND))
clf = XGBClassifier(max_depth=MAX_DEPTH,
random_state=42,
n_estimators=NUM_ROUND,
min_child_weight=18,
subsample=0.9,
colsample_bytree=0.6,
eta=ETA,
objective='binary:logistic',
verbosity=0,
n_thread=n_thread,
**param)
verbosity = True
if is_use_gpu or is_colab_cpu:
verbosity = False
clf.fit(tr_input_arr, tr_angle_arr, eval_set=watchlist, verbose=verbosity)
process_time = time.time() - start
logfile_writeln_tr("excecution time of training: " + str(process_time))
clf.save_model('./xgb.model')
booster = clf.get_booster()
booster.dump_model('./xgb_model.raw.txt')
eval_result_dic = clf.evals_result()
for ii in range(len(eval_result_dic['validation_0']['error'])):
if VALIDATION_DATA_RATIO != 0.0:
logfile_writeln_tr(
str(ii) + "," +
str(eval_result_dic['validation_0']['error'][ii]) + "," +
str(eval_result_dic['validation_1']['error'][ii]))
else:
logfile_writeln_tr(
str(ii) + "," +
str(eval_result_dic['validation_0']['error'][ii]))
# Feature Importance
fti = clf.feature_importances_
logfile_writeln_tr('Feature Importances:')
for i, feat in enumerate(FEATURE_NAMES):
logfile_writeln_tr('\t{0:20s} : {1:>.6f}'.format(feat, fti[i]))
log_fd_tr.flush()
log_fd_tr.close()
print("finished training and saved model.")
print("최종 정답률 : ", r2_score(y_test, y_predict))
score = accuracy_score(y_test, y_predict)
model = model.best_estimator_
thresholds = np.sort(model.feature_importances_)
print(thresholds)
n = 0
score = 0
for thresh in thresholds:
selection = SelectFromModel(model, threshold=thresh, prefit=True)
select_x_train = selection.transform(x_train)
selection_model = XGBClassifier(n_jobs=-1)
selection_model.fit(select_x_train, y_train)
select_x_test = selection.transform(x_test)
y_predict = selection_model.predict(select_x_test)
acc = accuracy_score(y_test, y_predict)
if acc * 100.0 > score:
n = select_x_train.shape[1]
score = acc * 100.0
L_selection = selection
selection_model.save_model("./save/xgb_save/ml37_3_cancer.xgb.model")
print("Thresh=%.3f, n=%d, acc: %.2f%%" %
(thresh, select_x_train.shape[1], score * 100.0))
def tree_optimization(self, learning_rate, gamma, max_depth, subsample,
reg_lambda, num_parallel_tree, min_child_weight):
x_train, x_eval, y_train, y_eval = train_test_split(
self.features,
self.labels,
test_size=0.2,
shuffle=True,
stratify=self.labels,
random_state=self.seed)
x_train, x_test, y_train, y_test = train_test_split(
x_train,
y_train,
test_size=0.2,
shuffle=True,
stratify=y_train,
random_state=self.seed)
booster_params = {
'n_estimators': 500,
'learning_rate': learning_rate,
'gamma': gamma,
'max_depth': int(np.around(max_depth)),
'subsample': subsample,
'sampling_method': 'gradient_based',
'reg_lambda': reg_lambda,
'min_child_weight': int(np.around(min_child_weight)),
'num_parallel_tree': int(np.around(num_parallel_tree)),
'objective': 'binary:logistic',
'verbosity': 1,
'max_delta_step ': 1
}
print("generating model")
model = XGBClassifier(**booster_params)
model.fit(X=x_train,
y=y_train,
eval_set=[(x_test, y_test)],
eval_metric=f1_eval,
early_stopping_rounds=25)
preds = model.predict(x_eval)
current_score = f1_score(y_eval, preds)
print(f"eval ROC score: {current_score}")
if current_score > self.best_score:
print(f"Score increased to {current_score} from {self.best_score}")
self.best_score = current_score
sub_pred = model.predict(self.pred)
model.save_model(
f'logs/f1_{self.run_num}_{self.best_score}_{self.tree_no}_{self.model_type}_thresh.model'
)
savetxt(
f'logs/{self.run_num}_{self.tree_no}_{self.model_type}_preds.txt',
sub_pred,
delimiter=',')
booster_params['f1'] = current_score
booster_params['tree_no'] = self.tree_no
record_history(booster_params)
self.tree_no += 1
return current_score
test_size=0.2,
random_state=24)
x_train = x_train.as_matrix()
y_train = y_train.as_matrix()
x_valid = x_valid.as_matrix()
y_valid = y_valid.as_matrix()
clf = XGBClassifier(booster='gbtree',
nthread=4,
learning_rate=0.1,
min_child_weight=1,
max_depth=5,
subsample=0.8,
colsample_bytree=0.8,
scale_pos_weight=1,
objective='binary:logistic',
seed=24)
eval_set = [(x_valid, y_valid)]
clf.fit(x_train,
y_train,
early_stopping_rounds=100,
eval_metric='logloss',
eval_set=eval_set,
verbose=True)
clf.save_model("./model/xgboost.model")
pre = clf.predict(x_valid)
pro = clf.predict_proba(x_valid)[:, 1]
print("Accuracy: %f" % accuracy_score(y_valid, pre))
print("AUC Score: %f" % roc_auc_score(y_valid, pro))
def train_xgboost(data: IrisData, artifacts_folder: str):
clf = XGBClassifier()
clf.fit(data.X, data.y)
clf.save_model(f"{artifacts_folder}/{XGBoostFolder}/model.bst")
# rmse, mae, logloss, error(설명 error가 accuracy), auc(설명 accuracy친구)
results = model.evals_result()
# print("eval's results : ", results)
# print("r2 Score : %.2f%%:" %(r2*100.0))
y_pred = model.predict(x_test)
acc = accuracy_score(y_pred, y_test)
print("acc : ", acc)
#####################################################################################################
# import pickle # 파이썬에서 제공한다.
# from joblib import dump, load
# import joblib
# pickle.dump(model, open("./model/xgb_save/cancer.pickle.dat", "wb")) # wb형식으로 저장하겠다.
# joblib.dump(model, "./model/xgb_save/cancer.joblib.dat")
model.save_model("./model/xgb_save/cancer.xgb.model")
print("저장됬다.")
# model2 = pickle.load(open("./model/xgb_save/cancer.pickle.dat", "rb"))
# model2 = joblib.load("./model/xgb_save/cancer.joblib.dat")
model2 = XGBClassifier()
model2.load_model("./model/xgb_save/cancer.xgb.model")
print('불러왔다.')
y_pred = model2.predict(x_test)
acc = accuracy_score(y_pred, y_test)
print("acc : ", acc)
class Classifier:
# for initializing train and test sets, classifier and accuracy score
# Change method to gpu_hist if you want xgboost to run on a GPU
def __init__(self,
params={
'objective': 'reg:squarederror',
'verbosity': 0
}):
self.X_train = []
self.X_labels = []
self.test = []
self.test_labels = []
self.model = XGBClassifier(**params)
self.prediction = 0
self.error = 0
def size(self):
if isinstance(self.X_train, np.ndarray):
return self.X_train.size
return len(self.X_train)
# adding the data points
def input_train(self, features, feature):
if isinstance(self.X_train, np.ndarray) and self.X_train.size > 0:
self.X_train = self.X_train.tolist()
self.X_labels = self.X_labels.tolist()
self.X_train.append(features)
self.X_labels.append(feature)
# train the data
def train(self):
self.X_train = np.asarray(self.X_train)
self.X_labels = np.asarray(self.X_labels)
self.model.fit(self.X_train, self.X_labels)
def train_eval(self, metric='error'):
self.X_train = np.asarray(self.X_train)
self.X_labels = np.asarray(self.X_labels)
X_train, X_test, y_train, y_test = train_test_split(self.X_train,
self.X_labels,
test_size=0.33)
self.model.fit(X_train,
y_train,
eval_set=[(X_train, y_train), (X_test, y_test)],
eval_metric=metric)
evals_result = self.model.evals_result()
if metric == 'error':
validations = []
for val in evals_result.values():
lst = val.get("error")
validations.append(sum(lst) / len(lst))
return 1 - (sum(validations) / len(validations))
else:
validations = []
for val in evals_result.values():
lst = val.get(metric)
validations.append(lst[-1])
return validations
# input test labels if you want to check accuracy
def label(self, label):
self.test_labels.append(label)
def input_test(self, features):
if isinstance(self.test, np.ndarray) and self.test.size > 0:
self.test = self.test.tolist()
self.test.append(features)
# test data
def predict(self):
if not isinstance(self.test, np.ndarray):
self.test = np.asarray(self.test)
self.prediction = self.model.predict(self.test)
return self.prediction
def predict_proba(self):
if not isinstance(self.test, np.ndarray):
self.test = np.asarray(self.test)
self.prediction = self.model.predict_proba(self.test)
return self.prediction
# if you have the test labels you can check the error rate (you want error close to 0)
def check_error(self):
self.test_labels = np.asarray(self.test_labels)
self.error = metrics.mean_absolute_error(self.test_labels,
self.prediction)
return self.error
# save classifier
def save_classifier(self, file):
self.model.save_model(file)
# open saved classifier
def open_classifier(self, file):
self.model.load_model(file)
# removes all training data
def clean_train(self):
self.X_train = []
self.X_labels = []
# removes all testing data
def clean_test(self):
self.test = []
self.test_labels = []
class Xgboost(object):
def __init__(self,
task="cla",
module_type="performance",
compute_task="cpu",
**params):
"""
:param task:
:param module_type:
:param compute_task:
:param params:
"""
assert task in ["cla", "reg"]
assert module_type in ["debug", "performance", "balance"]
assert compute_task in ["cpu", "gpu"]
self.task = task
self.module_type = module_type # 模块
if self.module_type == "debug":
params["n_jos"] = 1
elif self.module_type == "performance":
params["n_jos"] = cpu_count() # cpu核心数
else: # 性能模式
params["n_jos"] = cpu_count() // 2
self.compute_task = compute_task
if self.compute_task == "gpu": # 使用gpu
params["tree_method"] = "gpu_hist"
else: # 默认cpu
params["tree_method"] = "hist" # 使用的cpu
if self.task == "reg": # 做回归任务
self.model = XGBRegressor(
learning_rate=params.get("learning_rate", 0.3),
n_estimators=params.get("n_estimators", 100), # 树的个数100,即代数
max_depth=params.get("max_depth", 6), # 树的深度
min_child_weight=params.get("min_child_weight", 1), # 叶子节点最小权重
n_jobs=params.get("n_jos", None), # 线程数
gamma=params.get("gamma", 0), # 惩罚项中叶子节点个数前的参数
reg_lambda=params.get("lambda", 1), # lambda
reg_alpha=params.get("alpha", 0),
tree_method=params.get("tree_method", "auto"),
subsample=params.get("subsample", 1), # 随机选择100%样本建立决策树
colsample_bytree=1, # 随机选择80%特征建立决策树
objective=params.get("objective",
"reg:squarederror"), # 指定损失函数
# num_class=params.get("num_class", 2), # 不指定即为2分类
booster=params.get("booster", "gbtree"), # 使用的提升器
scale_pos_weight=1, # 解决样本不平衡问题
random_state=27, # 随机数
)
else: # 做的分类任务
self.model = XGBClassifier(
learning_rate=params.get("learning_rate", 0.3),
n_estimators=params.get("n_estimators", 100), # 树的个数100,即代数
max_depth=params.get("max_depth", 6), # 树的深度
min_child_weight=params.get("min_child_weight", 1), # 叶子节点最小权重
n_jobs=params.get("n_jos", None), # 线程数
gamma=params.get("gamma", 0), # 惩罚项中叶子节点个数前的参数
reg_lambda=params.get("lambda", 1), # lambda
reg_alpha=params.get("alpha", 0),
tree_method=params.get("tree_method", "auto"), # 树方法， 默认为auto
subsample=params.get("subsample", 1), # 随机选择100%样本建立决策树
colsample_bytree=1, # 随机选择80%特征建立决策树
objective=params.get("objective", "multi:softmax"),
# 指定损失函数 # 'binary:logistic 二分类交叉上
# num_class=params.get("num_class", 2), # 不指定即为2分类
booster=params.get("booster", "gbtree"), # 使用的提升器
scale_pos_weight=1, # 解决样本不平衡问题
random_state=27, # 随机数
)
"""
目标函数类型
具体查看 https://xgboost.readthedocs.io/en/latest/parameter.html
obejctive: 默认 reg:squarederror:
reg:squarederror: #回归平方误差
reg:squaredlogerror # 上述误差上取对数
reg:logistic logistic regression
reg:logistic 逻辑回归
binary:logistic 逻辑回归二分类， 输出为概率值
binary:logitraw 逻辑回归 2分类，输出为logits之前的得分
binary:hinge 用于二元分类的铰链损失。这使得预测为0或1，而不是产生概率。
multi:softmax: 多分类，需要指定num_class的类别
multi:softprob: 输出为概率 ndata*nclass 的矩阵，即，每行数据为分属类别的概率
"""
def train(self,
x_train,
y_train=None,
sample_weight=None,
base_margin=None,
eval_set=None,
eval_metric=None,
early_stopping_rounds=None,
verbose=True,
sample_weight_eval_set=None):
# print(self.model)
"""
:param x_train: 回归中，使用特征矩阵， array
:param y_train: 标签 array
:param eval_metric
:return:
"""
# 默认开启过早停止
# eval_metric in ["rmse","rmsle","mae","logloss","error","error@t", "merror","mlogloss","auc","aucpr",
# "ndcg","map","ndcg@n", "map@n","ndcg-", "map-", "ndcg@n-", "map@n-","poisson-nloglik",
# "gamma-nloglik","cox-nloglik","gamma-deviance","tweedie-nloglik","aft-nloglik"]
# eval_metric 参数可为字符串， 也可以是列表字符串的形式
if eval_metric: # 若需要使用评估模型模式，
assert eval_set # 要确保 测试集是存在的。
self.model.fit(X=x_train,
y=y_train,
sample_weight=sample_weight,
base_margin=base_margin,
eval_set=eval_set,
eval_metric=eval_metric,
early_stopping_rounds=early_stopping_rounds,
verbose=verbose,
sample_weight_eval_set=sample_weight_eval_set)
# early_stopping_rounds=10 过早停止的条件 # 默认使用值为10
# verbose=True # 是否开启冗余
def plot_loss(self): # 绘制loss
result = self.model.evals_result() #获取模型结果
epochs = len(result["validation_0"]["rmse"])
x_axis = range(0, epochs)
# 绘制loss曲线图
figure, ax = plt.subplots()
ax.plot(x_axis, result["validation_0"]["rmse"], label="Train")
ax.plot(x_axis, result["validation_1"]["rmse"], label="Test")
ax.legend()
plt.ylabel("loss")
plt.title("Xgboost Log Loss")
plt.show()
def predict(self, x_test):
"""
:param x_test: #使用np.array、scipy.sparse 用于预测
:return:
"""
my_pred = self.model.predict(data=x_test,
output_margin=False,
validate_features=True,
base_margin=None)
return my_pred
def plt_importance(self, figure_path=None, ifsave=True): # 绘制重要性特征
"""
:param figure_path: 图片保存路径
:param ifsave: 是否保存图片
:return:
"""
# 绘制特征重要性
fig, ax = plt.subplots(figsize=(15, 15))
plot_importance(self.model, height=0.5, ax=ax,
max_num_features=64) # 最多绘制64个特征
if ifsave:
if not figure_path:
plt.savefig(
"../model/XGBboost_model/Xgboost_featute_importance_before.png"
)
else:
plt.savefig(figure_path)
plt.show() # 显示图片
def _plt_importance_v1(self,
columns_name,
figure_path=None,
ifsave=True): # 绘制重要性特征，使用实际的列名进行替换
fig, ax = plt.subplots(figsize=(15, 15))
plot_importance_v1(self.model,
model_name="xgb",
columns_name=columns_name,
height=0.5,
ax=ax,
max_num_features=64) # 最多绘制64个特征
if ifsave:
if not figure_path:
plt.savefig(
"../model/XGBboost_model/Xgboost_featute_importance_after.png"
)
else:
plt.savefig(figure_path)
plt.show() # 显示图片
def plt_tree(self, num_tree): # 绘制树
"""
:param num_tree: 指定目标树的序号
:return:
"""
plot_tree(booster=self.model, num_trees=num_tree)
def plot_graphviz(self, num_tree): # 进行绘制graphviz
to_graphviz(self.model, num_trees=num_tree)
# 获取重要特征
def get_importance(self):
return self.model.feature_importances_
# 评估函数
def evaluate(self, y_test, my_pred, evalue_fun="mse"):
if evalue_fun == "acc": # 准确率 分类指标
result = accuracy_score(y_true=y_test, y_pred=my_pred)
print("accuarcy:%.2f" % (result * 100.0))
elif evalue_fun == "auc": # auc 值 分类指标
result = roc_auc_score(y_true=y_test, y_score=my_pred)
print("auc:%.2f" % (result))
elif evalue_fun == "mae": # 回归指标， 平均绝对误差
result = mean_absolute_error(y_true=y_test, y_pred=my_pred)
print("mae:%.2f" % (result))
elif evalue_fun == "median_ae": # 种植绝对误差 回归指标
result = median_absolute_error(y_true=y_test, y_pred=my_pred)
print("median_ae:%.2f" % (result))
elif evalue_fun == "r2_score": # R平方值 回归指标
result = r2_score(y_true=y_test, y_pred=my_pred)
print("r2_score:%.2f" % (result))
elif evalue_fun == "evs": # 回归反差， 回归指标
result = explained_variance_score(y_true=y_test, y_pred=my_pred)
print("explained_variance_score:%.2f" % (result))
elif evalue_fun == "aps": # 分类指标， 根据预测得分计算平均精度(AP)
result = average_precision_score(y_true=y_test,
y_score=my_pred,
average="maco",
sample_weight=None)
print("average_precision_score:%.2f" % (result))
elif evalue_fun == "bsl":
result = brier_score_loss(y_true=y_test,
y_prob=my_pred,
sample_weight=None,
pos_label=None)
print("brier_score_loss:%.2f" % (result))
elif evalue_fun == "cmt": #计算混淆矩阵来评估分类的准确性 分类指标
result = confusion_matrix(y_true=y_test,
y_pred=my_pred,
labels=None,
sample_weight=None)
print("confusion_matrix:%.2f" % (result))
elif evalue_fun == "f1_score": # f1 得分， 分类指标
result = f1_score(y_true=y_test,
y_pred=my_pred,
labels=None,
pos_label=1,
average="binary",
sample_weight=None) #F1值
print("f1_score:%.2f" % (result))
elif evalue_fun == "log_loss": # 交叉熵孙绍， 分类指标
result = log_loss(y_true=y_test,
y_pred=my_pred,
eps=1e-15,
normalize=True,
sample_weight=None,
labels=None)
print("log_loss:%.2f" % (result))
elif evalue_fun == "precision_score": # 查准率 分类指标
result = precision_score(y_true=y_test,
y_pred=my_pred,
labels=None,
pos_label=1,
average="binary")
print("precision_score:%.2f" % (result))
elif evalue_fun == "recall_score": # 查全绿 分类指标
result = recall_score(y_true=y_test,
y_pred=my_pred,
labels=None,
pos_label=1,
average="binary",
sample_weight=None)
print("recall_score:%.2f" % (result))
elif evalue_fun == "roc_auc_score": # 计算 roc 曲线下面的面积就是AUC值， 分类指标
result = roc_auc_score(y_true=y_test,
y_score=my_pred,
average="macro",
sample_weight=None)
print("roc_auc_score:%.2f" % (result))
elif evalue_fun == "roc_curve": # 计算PROC曲线的横轴坐标 分类指标
fpr, tpr, thresholds = roc_curve(y_true=y_test,
y_score=my_pred,
pos_label=None,
sample_weight=None,
drop_intermediate=True)
result = (fpr, tpr, thresholds)
else: # mse 参数 均方差， 回归指标
result = mean_squared_error(y_true=y_test, y_pred=my_pred)
print("mse：%.2f" % (result))
return result
def save_model(self, save_params): # 模型保存
self.model.save_model(
fname=save_params.get(
"fname",
"../model/XGBboost_model/XGboostmodel.model") # 保存的文件路径名字
# format=save_params.get("format", "cbm"), # 保存的数据格式
# pool=save_params.get("pool", None) # 训练使用的数据 模型保存成json格式，无需使用pool
)
