A Guide to Cross-Validation for AI - Avoiding Overfitting and Ensuring Generalizability

A Guide to Cross-Validation for AI - Avoiding Overfitting and Ensuring Generalizability

Molecular Imaging & Therapy via YouTube Direct link

selecting an approach: pros and cons

13 of 14

13 of 14

selecting an approach: pros and cons

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

A Guide to Cross-Validation for AI - Avoiding Overfitting and Ensuring Generalizability

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Introduction
  2. 2 Overfitting vs. generalizability
  3. 3 Pitfalls of using one-time split method
  4. 4 Pitfall #1: Non-representative test set
  5. 5 Pitfall #2: Tuning to the test set
  6. 6 Cross-validation
  7. 7 Important note: in CV we are testing pipeline, not a single model
  8. 8 K-fold, folded test set
  9. 9 K-fold, hold-out test-set
  10. 10 Nested cross-validation
  11. 11 leave-one-out
  12. 12 random sampling
  13. 13 selecting an approach: pros and cons
  14. 14 Final thoughts

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.