my_custom_scorer = make_scorer(my_pred, greater_is_better=True)
tpo = TPOTClassifier(generations=100,
verbosity=2,
population_size=100,
scoring='f1',
n_jobs=-1,
config_dict=tpot_config,
mutation_rate=0.9,
crossover_rate=0.1,
cv=5,
random_state=5)
tpo.fit(x_train, y_train)
evalution_model(tpo, x_train, y_train)
evalution_model(tpo, x_test, y_test)
evalution_model(tpo,
np.array(df_btest.drop("is_sucess_by_contract", axis=1)),
df_btest["is_sucess_by_contract"])
import datetime
i = datetime.datetime.today()
print(i)
s = str(i.month) + str(i.day) + str(i.hour) + str(i.minute)
model_name = "GM_export/main_new/" + "GM" + s + ".py"
tpo.export(model_name)
import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier
train_set=dtrain,
num_boost_round=100,
valid_sets=[dtrain],
valid_names=None,
fobj=None,
feval=None,
init_model=None,
categorical_feature='auto',
early_stopping_rounds=None,
evals_result=evals_result,
verbose_eval=10,
keep_training_booster=False,
callbacks=None,
)
print("================训练集================")
evalution_model(clfs, x_train, y_train)
print("================测试集================")
evalution_model(clfs, x_test, y_test)
print("===========b_test===================")
evalution_model(clfs, df_xbtest, df_ybtest)
# 查找每棵树叶子节点
y_pred_train = clfs.predict(x_train, pred_leaf=True)
y_pred_test = clfs.predict(x_test, pred_leaf=True)
y_pred_btest = clfs.predict(df_xbtest, pred_leaf=True)
# onehot编码
enc = OneHotEncoder()
enc.fit(y_pred_train)
train_encode = np.array(enc.transform(y_pred_train).toarray())
test_encode = np.array(enc.transform(y_pred_test).toarray())
btest_encode = np.array(enc.transform(y_pred_btest).toarray())
# model.save('models/deep_model/lstm_0228.h5')
# model = load_model('models/deep_model/lstm_0228.h5')
# classes = model.predict_classes(xt)
# acc = np_utils.accuracy(classes, yt)
# score, acc = model.evaluate(xt, yt, batch_size=16, verbose=1)
score, acc = model.evaluate(x_test, y_test, verbose=1)
print('Test score:', score)
print('Test accuracy:', acc)
y_pred_train = model.predict(x_train)
y_pred_test = model.predict(x_test)
y_pred_btest = model.predict(x_btest)
print("================训练集================")
evalution_model(model, x_train, y_train)
print("================测试集==============")
evalution_model(model, x_test, y_test)
print("================B测试集==============")
evalution_model(model, x_btest, y_btest)
y_pred_train_prob = model.predict_proba(x_train)
y_pred_test_prob = model.predict_proba(x_test)
y_pred_btest_prob = model.predict_proba(x_btest)
x_trains["voice_prob"] = y_pred_train_prob
x_tests["voice_prob"] = y_pred_test_prob
x_btests["voice_prob"] = y_pred_btest_prob
)
fg = rf_bo.max
clfs = RandomForestClassifier(
n_estimators=int(fg["params"]["n_estimators"]),
min_samples_split=int(fg["params"]["min_samples_split"]),
max_features=int(fg["params"]["max_features"]),
# float
max_depth=int(fg["params"]["max_depth"]),
random_state=5,
class_weight={1: 1.46})
clfs.fit(x_train, y_train)
print("================训练集================")
evalution_model(clfs, x_train, y_train)
print("================测试集================")
evalution_model(clfs, x_test, y_test)
print("===========b_test===================")
evalution_model(clfs, df_btest.drop("is_pigeon", axis=1),
df_btest["is_pigeon"])
#catboost基模型
# from catboost import CatBoostClassifier, CatBoostRegressor, Pool
#
#
# def catboost_cv(n_estimators, learning_rate, depth, l2_leaf_reg):
# clf = CatBoostClassifier(
# n_estimators=int(n_estimators),
# learning_rate=learning_rate,
# depth=int(depth),