EVRI-thing You Need to Know About Cross-Validation

Cross-validation is a ubiquitous but often mis-used machine learning procedure. Eigenvector’s Barry M. Wise takes an in-depth review of this important methodology for establishing model meta-parameters and estimating predictive performance.

Table of Contents

00:00 – Introduction and Outline

02:40 – Basics of Cross Validation

11:20 – Ways to Split the Data

26:42 –  Things to Consider when Choosing a CV Method

44:35 –  Setting up custom CV

53:42 – Cross Validation in PCA

55:32 – How to Make Cross-validation go Bad

1:00:22 – Variations and Related Methods

1:01:54 – Conclusions