ROC and AUC, Clearly Explained!

35 comentários em “ROC and AUC, Clearly Explained!”

1. NOTE: This revised video corrects three minor errors (that, fortunately, did not affect the main idea) found by vigilant viewers in the original. Other than that, it's the same as before. That said, there is a minor error at 12:00 . The confusion matrix should be TP = 3, FP = 2, FN = 1, TN = 2. The displayed matrix should be for the next point.

2. I've just upvoted the intro song. That's it. Oh wait, the video also explains the ROC and AUC? that's great!

3. "AUC" is probably the most undescriptive acronym I have never heard of. Who comes up with these names? Well clearly they were gifted more mathematically than lingually since the idea is pretty cool.

4. 7:32 confusion matrix

5. Love the content, but having ads every 2 minutes is a bit too much. If you could push them towards the end would be much better

6. Thanks

7. Honestly I had zero knowledge of any theory or terminology or even what this subject is and I’m baked as fuck and I now have a basic understanding of this concept. A fine job sir

8. Such a great explanation! Simple text, simple images, BAAAM!

9. bam!

10. Great video and great explanations! The threshold setting is the key point.😃

11. helped me so much, thank you

12. Thank you so much for this amazing explanation. This is the first time I actually understood the idea behind AUC and ROC.

13. You are an awesome man. Thanks…

14. bamm

15. Don't know if I'm the only one, I love the content, it's very high-quality and well-explained. Just the opening guitar music makes me uncomfortable. I would appreciate shorter videos with clear-cut content introduction.

16. i like the BAAAAA~~~~MMM

17. Thank you for this!

18. excellent as always

19. i still don't get how the AUC is 0.9

20. It's confusing…no pun intended…especially when dealing with health care or anything else where a score of 1, is actually the negative outcome or the outcome you do not want to occur. So in this dataset, "is obese" is represented by 0 and "is not obese" is 1?

21. guaranteed, you will be "BAMED" after watching this presentation.

22. Thank you for the wonderful work Josh! Just bought a StatQuest Triple BAM shirt. Is the threshold arbitrarily set to change by increments ? If so, what is the typical increments for thresholds?

23. I was struggling to understand this until the double BAM

24. Thanks for the explanation.
    I am planning to go throw all the videos on the machin leaning playlist as it seems to have a lot of valuable informations.
    I can take it in 3 or 4 days but I want to implement and test these concepts using real world machine learning dataset and Scikit-Learn
    This will make it takes about 8 to 10 days! Excited to achieve this goal 😊

25. Maybe I have a stupid quastion but… I'm wondering how do you colorize the area under the curve for ROC in R? I've managed to take only the value for auc (with [email protected][[1]]). And very useful video by the way 🙂

26. This is an amazing explamation. Thank you for that. I have a question though. When we talk about object detection, which is a classification problem, hwo do we deal with true negatives? It could be a very large number. So do we use the false positive rate for object detection like YOLO or do we use something else?

27. Such an awesome video. ur a life saver

28. Бам) спасибо, очень интересно

29. Best explanation ever!!!!

30. great explaining thank you

31. BAM!

32. Thank you for explaining it so clearly!

33. Thank you so much. Your video is very useful!

34. Thank you, Josh. Your videos are great. Do you by any chance have something on Cohen's Kappa and TSS? Cheers.

35. This is the difficult one, but you made a great explanation. That is marvelous.

Os comentários estão fechados.