# Use test data as train: does it make sense?

There is a classification problem(two classes). We have train data, for which we know class labels and we have test data.

Imagine, that you have created model that with good accuracy(~95%) make predictions and we know that we are not overfitted.

If we make prediction on test data, extract objects for which we sure in class label(for example, predict_proba higher than 90%) and add this objects to train data.

Does this tactic make any sense?

• You have labeled training data, but unlabeled test data? How exactly are you testing? Is this a Kaggle competition or similar? – Neil Slater Mar 28 '16 at 20:35
• You will have to elaborate more on this -- What is your rationale for adding records to the training data? What do you intend to do after adding these records to the training data? – Vishal Mar 29 '16 at 0:22