Preprocessing that fits on train only on the How to Train Models track. Means, vocabularies, and imputation values are parameters. Fit them on train. Apply them to validation and test. Refit them inside each cross-validation fold.
This lesson assumes you already worked through Freeze the split.
The idea in practice
Store the fitted preprocessor with the model so production uses the same numbers.
A concrete check
goal = {
'track': 'How to Train Models',
'lesson': 'Preprocessing that fits on train only',
}
checks = [
'input available at decision time',
'score matches the real decision',
'failure case written down',
]
print(goal['lesson'])
for item in checks:
print('-', item)
Run the sketch locally if you have Python. The printout is a reminder of the checks, not a trained model. Replace the strings with the real inputs from your own example before you treat it as a design.
What usually goes wrong
Computing the mean on all rows, then splitting. When this happens, stop adding parameters or tools. Fix the check, the data, or the permission, then run the same example again.
What to write down
- The input you are allowed to use at decision time.
- The output and the score or pass rule.
- One failure you will test on purpose.
- What you will not claim the system can do.
Practice
List two statistics that must be fit on train only.
Self-check
- Say Preprocessing that fits on train only in one sentence that mentions an input and an output.
- Name the failure mode in this lesson and the check that would catch it.
Done when: you can explain this lesson without the page open, and you have a written failure case.