Code and instructions for techniques to properly validate your machine learning models.
In this repo, you will find the code and instructions for this article. It is advised to read through the article whilst coding along using the Validation.ipynb notebook.
I believe that one of the most underrated aspects of creating your Machine Learning Model is thorough validation. Using proper validation techniques helps you understand your model, but most importantly, estimate an unbiased generalization performance.
This repo and the corresponding article describe several methods for advanced validation of machine learning algorithms:
- Train/test split
- k-Fold CV
- Leave-one-out CV
- Leave-one-group-out CV
- Time-series CV
- Nested CV
- Wilcoxon signed-rank test
- McNemar's test
- 5x2CV paired t-test
- 5x2CV combined F test