Splitting the Data: Train, Validation and Test

56 min
Block 7 — Train, validation and test
Objective
establish why an error measured on the data that built a model carries no information about generalization; separate the roles of the three sets; distinguish model selection from model assessment and show why one set cannot perform both; master train_test_split and the semantics of random_state; size the partitions from the precision required on the announced figure; quantify the variance of a single split and reduce it with cross-validation; know the splitter family and the bias–variance trade-off of k; run a nested cross-validation; reserve hold-out data for threshold tuning and calibration; and recognize a contaminated test set.
Estimated duration
70 minutes
Prerequisites
chapters 005 (dataset structure), 008 (algorithm and model), 009 (parameters and hyperparameters), 012 (life cycle, stage 8) and 025 (feature selection)
Associated quizzes
026.1-quiz-biased-training-error.md to 026.8-quiz-nested-cv-and-contamination.md

Preview — the rest of the lesson is for enrolled readers.

Already enrolled with a code?

Your access is tied to your account, not to this link. Sign in with the same email you used in class: your course is waiting, no need to enter the code again.

Sign inNo account yet? Create one
This lesson is part of the “Splitting the Data: Train, Validation, Test” module

The first modules of the course are open to everyone. For the rest you have three options: buy this course once and for all, subscribe, or enter the code handed out in class.

Are you a student on this course?

The code is tied to your account: sign in or create an account and it will be applied automatically when you come back.

No account yet? Create one