What is the normal time for running a voting classifier from sklearn?

Question

I am new to ensemble learning and methods and I have built the following model using sklearn

std = RobustScaler()
std.fit(train_feats)
train_feats = std.transform(train_feats)
val_feats = std.transform(val_feats)

# define base learners
# XGB classifier
xgb_classifier = xgb.XGBClassifier(objective='binary:logistic',
                                  learning_rate=0.1,
                                  n_estimators=10,
                                  max_depth=1,
                                  subsample=0.4,
                                  random_state=234)

# SVM
svm_classifier = SVC(gamma=0.1,
                    C=0.1,
                    kernel='poly',
                    degree=3,
                    coef0=10.0,
                    probability=True)

# random forest classifier
rf_classifier = RandomForestClassifier(n_estimators=10,
                            max_features="sqrt",
                            criterion='entropy',
                            class_weight='balanced')
# Define meta-learn
voting_clf = VotingClassifier([("xgb", xgb_classifier),
                               ("svm", svm_classifier),
                               ("rf",  rf_classifier)],
                             voting="soft",
                             flatten_transform=True)

voting_clf.fit(train_feats, train_labels)

The model has been running for 5 hours. The shape of train feats is: (18000, 29). Is it normal for the voting classifier to be running for 5 hours with no sign of ever stopping? Is there a bug here? I don't want to stop the training and re-run unless I know things are wrong and there is a bug.

I was curious if there is a bug that is slowing the training time or is it generally like this, that takes long time to be trained?

Answer 1

It seems that SVC runs in quadratic time as mentioned in this post. So removing the SVM solved the issue for now.