def objective_svm(x):
C_in, gamma_in = x[0]**2, x[1]**2
svm_classifier = SVC(C=C_in,
cache_size=200,
class_weight=None,
coef0=0.0,
decision_function_shape='ovr',
degree=3,
gamma=gamma_in,
kernel='rbf',
max_iter=-1,
probability=False,
random_state=None,
shrinking=True,
tol=0.001,
verbose=False)
k = 5
kf = StratifiedKFold(k)
i = 0
score = 0
for train_index, test_index in kf.split(X_train, Y_train):
svm_classifier.fit(X_train[train_index], Y_train[train_index])
Y_pred = svm_classifier.predict(X_train[test_index])
score += f1_score(Y_train[test_index], Y_pred)
i += 1
return score
lgb_eval = lgb.Dataset(X_train[test_index],
Y_train[test_index],
reference=lgb_train)
gbm = lgb.train(parameters,
train_set=lgb_train,
num_boost_round=100,
valid_sets=lgb_eval,
verbose_eval=40,
feval=f1_score_lgbm)
res = gbm.predict(X_test)
Y_pred = gbm.predict(X_train[test_index])
Y_pred_train = gbm.predict(X_train[train_index])
predictions[:, i] = res
predictions_train[test_index] = Y_pred
print("train: " +
str(f1_score(Y_train[train_index], Y_pred_train.round())))
print("test: " + str(f1_score(Y_train[test_index], Y_pred.round())))
i += 1
# save submission file
Y_test = (np.sum(predictions, axis=1) > 2.5).astype(int)
submission = pd.DataFrame(Y_test)
submission.to_csv(path_or_buf=path_to_submissions +
"-".join(my_features_acronym) + "lgbm" + ".csv",
index=True,
index_label="id",
header=["category"])
# save probabilities for stacking
stacking_logits_test = np.sum(predictions, axis=1)
stacking_test = pd.DataFrame(stacking_logits_test)
kf = KFold(k)
predictions = np.zeros((X_test.shape[0], k))
predictions_test = np.zeros((X_test.shape[0], k))
predictions_train = np.zeros(X_train.shape[0])
i = 0
# for each fold store predictions on test set and print validation results
test_score = 0.0
for train_index, test_index in kf.split(X_train, Y_train):
RF.fit(X_train[train_index], Y_train[train_index])
Y_pred = RF.predict(X_train[test_index])
Y_pred_train = RF.predict(X_train[train_index])
predictions[:, i] = RF.predict(X_test)
predictions_test[:, i] = RF.predict_proba(X_test)[:, 1]
predictions_train[test_index] = RF.predict_proba(X_train[test_index])[:, 1]
current_test_score = f1_score(Y_train[test_index], Y_pred)
test_score += current_test_score
print("train: " + str(f1_score(Y_train[train_index], Y_pred_train)))
print("test: " + str(current_test_score))
i += 1
print("CV test score: " + str(test_score / k))
# save submission file
Y_test = (np.sum(predictions, axis=1) > 2.5).astype(int)
submission = pd.DataFrame(Y_test)
submission.to_csv(
path_or_buf=path_to_submissions + "-".join(my_features_acronym) + "RF.csv",
index=True,
index_label="id",
header=["category"]
)
print("parameters:")
print(parameters)
print("cross validation:")
LogReg = LogisticRegressionCV(max_iter=parameters['max_iter'],
tol=parameters['tol'],
penalty=parameters['penalty'])
k = 5
kf = KFold(k)
predictions = np.zeros((X_test.shape[0], k))
i = 0
for train_index, test_index in kf.split(X_train, Y_train):
LogReg.fit(X_train[train_index], Y_train[train_index])
Y_pred = LogReg.predict(X_train[test_index])
Y_pred_train = LogReg.predict(X_train[train_index])
predictions[:, i] = LogReg.predict(X_test)
print("train: " + str(f1_score(Y_train[train_index], Y_pred_train)))
print("test: " + str(f1_score(Y_train[test_index], Y_pred)))
i += 1
Y_test = (np.sum(predictions, axis=1) > 2.5).astype(int)
# submission = pd.DataFrame(Y_test)
# submission.to_csv(
# path_or_buf=path_to_submissions+"-".join(my_features_string)+"LogReg.csv",
# index=True,
# index_label="id",
# header=["category"]
# )
n_estimators=parameters["n_estimators"],
criterion=parameters["criterion"],
max_depth=parameters["max_depth"],
min_samples_leaf=parameters["min_samples_leaf"],
bootstrap=parameters["bootstrap"],
n_jobs=parameters["n_jobs"])
k = 2
kf = KFold(k)
train_score = 0.0
test_score = 0.0
for train_index, test_index in kf.split(X_train, Y_train):
RF.fit(X_train[train_index], Y_train[train_index])
Y_pred = RF.predict(X_train[test_index])
Y_pred_train = RF.predict(X_train[train_index])
train_score += f1_score(Y_train[train_index], Y_pred_train)
test_score += f1_score(Y_train[test_index], Y_pred)
train_score /= k
test_score /= k
if test_score > best_test_score:
best_index = i
best_train_score = train_score
best_test_score = test_score
print("train score: " + str(train_score))
print("test score: " + str(test_score))
print("")
print("for this round, the best feature was " +