ROC and AUC, Clearly Explained!
Vložit
- čas přidán 2. 06. 2024
- ROC (Receiver Operator Characteristic) graphs and AUC (the area under the curve), are useful for consolidating the information from a ton of confusion matrices into a single, easy to interpret graph. This video walks you through how to create and interpret ROC graphs step-by-step. We then show how the AUC can be used to compare classification methods and, lastly, we talk about what to do when your data isn't as warm and fuzzy as it should be.
NOTE: This is the 2019.07.11 revision of a video published earlier.
NOTE: This video assumes you already know about
Confusion Matrices...
• Machine Learning Funda...
...Sensitivity and Specificity...
• Machine Learning Funda...
...and the example I work through is based on Logistic Regression, so it would help to understand the basics of that as well:
• StatQuest: Logistic Re...
For a complete index of all the StatQuest videos, check out:
statquest.org/video-index/
If you'd like to support StatQuest, please consider...
Buying The StatQuest Illustrated Guide to Machine Learning!!!
PDF - statquest.gumroad.com/l/wvtmc
Paperback - www.amazon.com/dp/B09ZCKR4H6
Kindle eBook - www.amazon.com/dp/B09ZG79HXC
Patreon: / statquest
...or...
CZcams Membership: / @statquest
...a cool StatQuest t-shirt or sweatshirt:
shop.spreadshirt.com/statques...
...buying one or two of my songs (or go large and get a whole album!)
joshuastarmer.bandcamp.com/
...or just donating to StatQuest!
www.paypal.me/statquest
Lastly, if you want to keep up with me as I research and create new StatQuests, follow me on twitter:
/ joshuastarmer
0:00 Awesome song and introduction
0:48 Classifying samples with logistic regression
4:03 Creating a confusion matrices for different thresholds
7:12 ROC is an alternative to tons of confusion matrices
13:44 AUC to compare different models
14:28 False Positive Rate vs Precision (Precision Recall Graphs)
15:38 Summary of concepts
Correction:
12:00 The confusion matrix should be TP = 3, FP = 2, FN = 1, TN = 2. The displayed matrix should be for the next point.
#statquest #ROC #AUC
Correction:
12:00 The confusion matrix should be TP = 3, FP = 2, FN = 1, TN = 2. The displayed matrix should be for the next point.
Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/
Professor please make videos on Gaussian mixture models, quantum Gaussian mixture models and deep learning related models like cnn, rnn, lstm etc.
@@GauravSharma-ui4yd I would very much like that too! Your videos have really helped me in my machine learning and stat courses! Thank you :)
Josh, you are a life safer! I can't express how much I appreciate your content and overall approach.
Lolol😭😍😭😭😭😁
replace ebola with corona
There’s a special place in heaven for those who make complicated things seem easy.
Thanks for the explanation
Thank you so much! :)
@@statquest You are simply awesome!
@@shilpishankermishra Thanks! :)
Awsome squared
@skandar This is the best response so far! Triple BAM!!!
dude you ROC!!
I love it! :)
Best comment
😂
Double Bam!
@@Phantom_Blox Dude you rock! Ay, you see what I did there? 😆
I am binge-watching statquest. Josh Starmer has lifted the bar of teaching too high. He is singing, he is playing guitar, and explaining technical concepts in a way that anyone can understand. What a unique combination!!!
Thank you very much! And thank you for your support!
The funny "BAM" points exactly match the moments I feel something learned.
Thanks for the great visualization!
Awesome!!! :)
Now I finally know what to think whenever he says "BAM" XD
@@samarthsoni4267 It's kind of like catch phrases of kid shows. I love it.
double bam!!!!!!!!!
bro you just did what my lecturer couldnt do the entire semester, in 15 minutes. props.
Thanks! :)
May be he could not understand well
So your professor was teaching ROC for the entire semester
@@op711 yeah “bro”
@@op711 lol ;)
I am a data scientist and I have never really understood ROC/AUC curve fully before, but this video is so well that I immediately get it. Thank you OG!!!!
BAM! :)
Hey Albert, do you professionally work as a Data Scientist??
@@rajshah7734 yes
If I ever pass my PhD comprehensive exam, it will be because God guided me to find your videos :D Thank you man!
I hope it goes well! Let me know! :)
Lolol same here! I have my exam on Wednesday!
I cried about this for hours and watched this video and understood everything. You are amazing
Thank you!
This guy is no joke the best at taking concepts and making them easy to understand. Props to you and thank you for doing what you do!
Thank you very much! :)
13:48 "BAM?" I laughed a lot! 😂😂😂 You're one of the best teachers on YT, Josh! Thanks a lot!
Thank you!
I have never felt so lucky to find a CZcams channel before. Double BAM!
Hooray! :)
I'm studying for my machine learning finals and your video's (not only this one) help me allot. Thanks for taking the time to make these!
Good luck with your finals! :)
With this video my search for ROC-AUC stopped. Thanks a lot for awesome explantion
BAM! :)
Thank you! It is incredible how well you are able to explain complicate concepts so simply. This is by far one of the best channels in youtube.
Wow, thanks!
Great Josh, I really like the step-by-step explanation of so called "complicated" things
Thank you very much! :)
Studying ML and classification right now and this helped connect so many dots in a way that was easy and repeatable! Thank you so much for the time you take to make these videos.
Thank you very much! :)
The world needs people like you! I love your extremely simple and clear explanations, style and of course your voice
Thank you! 😃
This is a random comment from Japan.
Thank you for your amazing contents !!
Studying data science is super hard for me even in mother language,Japanese.
But your contents are awesome.
Your are a lifesaver for data-science-leaners all around the world!
Hooray! Thank you very much! :)
You are the lecturer no one has but everyone needs.
:)
I think this is the first time I trully understand the ROC and AUC. Really, thanks!
Hooray!!!
Man your channel is a hidden gem. Thank you for the time and effort you put into statquests!
Thank you!
Million BAM's , I can't thank you enough for being the teacher I never had!!!
Thanks!
you are the only reason I understand anything in machine learning
bam!
This is the clearest explanation of the ROC curve I've yet seen. Kudos to you for developing this!
Thank you very much! :)
This absolutely saved me as I finished a final project! Thank you so much for making these clear & helpful videos
Hooray and congratulations on finishing your final project!! BAM! :)
My God! If this is not THE BEST video about ML algorithms, then I don’t know which is! Thank you so much! Maximum share among classmates! Hope you dude always be happy!
Thank you for sharing! :)
If there is someone with the gift of teaching, it's you! Kudos!
Thank you!
I’m a guy from ROC (Taiwan) watching this ROC tutorial and understanding ROC. Thank you!
Rock on!
every time i watch his videos i felt very relaxed and feel like there is something heavenly , divine in the world
:)
I love this channel! It is fun and make dense subject way easier. Thank you for sharing high quality material with us!
Thank you very much! :)
I have been trying to understand this for soo long. Can't describe in words how amazingly simple you have made this. Thank you so much
Hooray! :)
I really enjoy the thought processes drawn and presented so vividly and clearly!!! Thank you so much for the video. Reading tons of materials but still couldn't understand but understood ROC and AUC the moment finished watching your video. Really appreciate that.
Glad it was helpful!
I've read so many Medium articles about ROC and AUC, but none describe it as good as your video!
Thanks!
Content = ok; Presentation = ok; Explanation = ok; Pronunciation = of => Perfect Video! Again, Salute from Brazil!
Thanks!
Thank you so much. Sometimes in advanced epi I feel just overwhelmed with all the details. Your videos really help keep the over all picture in mind. Love them!!
Hooray! Thank you very much! :)
My mind says ..."TRIPLE BAMMMMM !!!" and "HURRAAYY" after grasping this complicated concept. You are a GEM Josh. Thanks for explaining it so slow and so perfectly like water going inside my throat.
Thank you so much!!!! I'm glad the video was helpful. :)
you are simply out of the universe , thanks for making my and other learner life easy.
Thanks! :)
I've never understood the ROC and AUC curves until now...BAM!!! I was so impressed that I just joined the StatQuest membership, thanks Josh Starmer!
BAM! Thank you very much for your support!!! :)
This man deserves a nobel prize
Thanks!
Josh you nail stats in every video, great.
This confusion matrix evaluation will be completed if you add one video on precision, recall, f1 score and precison recall curve
Thanks and waiting for it.
And after this video I got all the answers after wandering one resource to another.
Thanks a lot for such amazing videos...🙏
Thanks!
Excellent! I was confused by a popular online interactive data science course in their way of explaining the ROC curves. Your explanation made it crystal clear. Thank you very much!
Hooray! Glad it was helpful!
Your videos are AMAZING! Thanks for taking the time to do them. You explain everything very clearly and interactively. Thank you!
Thank you!
OMG, I was trying so hard to understand this and finally I found this. U totally ROC!!!!
Awesome! Thank you! :)
7:05 "a confusingly large number of confusion matrices." Classic! :)
Thanks! :)
Some teachers make simple things complicated... you, however, make complicated things seem simple. Thanks so much!
Thank you! :)
Dear Josh Starmer,
Thank you for your insightful lecture on ROC and AUC. Your explanations were clear, and I gained a better understanding of these concepts. Your dedication to simplifying complex topics is commendable.
Thank you!
This was so helpful, could you also make a video about transformation functions like Box Cox and arcsine? Would greatly help. :)
Thank you, sir. You are a gentleman and a scholar.
Thanks! :)
Thank you so much! I didn’t quite understand what the points and the curve meant, and then... BAM! Insight/understanding came down from the heavens!
You seriously have ROCed the world of a young medical student
May your channel get more and more viewers and you get more and more love
Thank you very much! :)
I was drinking my morning coffee, took a sip and this is the first video of yours that I just clicked on, and that intro! bruv! took me back to days in undergrad struggling. Legit spew coffee on myself. Such a well explained video. My respects.
That's awesome! BAM! :)
I wouldn't be able to express how easy it was to understand through the explanation. I wish I could like the same video 1 million times.
Thank you very much. :)
This is just incredible
I have had a very unclear understanding of this for about 3 years now. This video just solved my problem in approx. 15 mins. Great job!
Thanks a lot for this!
Thank you!
This is literally the best and the most clear explanation to ROC and AUC ! Thanks for doing this ! Bam !
Thank you! :)
Thank you so much for all the BAMs and the easy explanation ! Love it :)
Glad you like them!
Congrats on these series of videos!! Im currently doing a compressed introduction course on ML methods in my university, and watching these videos in paralel really helps my understanding of the given concepts!! Really good summarize!!
Awesome! I'm glad to hear the videos help. :)
@@statquest They do! Very much so! :) Thank you for making them :)
the enthusiasm and dedication you put in videos make learning easy.......Thank you very much!!!!
Thank you! :)
This is the best explantion i have ever found online for roc. funny how many people complicate simple concepts by throwing in lots of jargons. the examples walkthrough made it super simple and intuitive to understand this. Triple BAM Indeed!
BAM!
I clicked on this video as it appeared on my feed thinking it was an old one (also, I needed to revise my ROC concepts). Little did I know that it was uploaded literally minutes ago. *When youtube recommendation engines read your mind, TRIPPLE BAAMMM*
That's totally awesome! :)
ohh god at last i understood this concept through ur vdo .thanx BAM!!!!!!!!
Thanks! :)
Such a lucid explanation of such a bootstrapped concept, which I had a feeling of having understood , but never clearly. Thanks
You are welcome!
The goat of ML explanations online. If only all teachers were half as passionate/explained half as well. Thanks for your amazing work.
Wow, thank you!
after watching this video i felt there is someone in the world who can save me, thanks to Joshua
:)
BAM!!! truely an amazing video
Many thanks........
Thank you! :)
dude, you keep things so intuitive, you have a gift, please keep teaching us
Thanks, will do!
Simplest possible explanation for a relatively complex topic. Hats off 👏
Thank you! :)
Arghhhhhh. Thank You for this. I was losing my mind already.
BAM! :)
Me *understand ROC and AUC because of watching this video*: TRIPLE BAM!!!
BAM! :)
I love the way this channel can grab my like 5 seconds into the video
bam!
best explanation of the ROC and AUC so far. Thank you for explaining so clearly
Wow, thanks!
immediate like for the intro !
Hooray! :)
14:25 DOUBLE BAAAAM!!! XD Because of those jokes I'll never forget what I've learned in this video. Great way to use humor, dude
Thank you! :)
TRIPLE BAM!!!
Hello Josh! you are an outstanding teacher! all your explanations are clear and are accessible to all, even to those who have little or no knowledge. your work is extremely important and I thank you wholeheartedly for sharing your great knowledge with us. It was a real pleasure to watch this video so well explained!!
Thank you very much! :)
I admire your videos. So clear reasonable explanation on the complicated topics. Worth thousands of written papers!
Thanks!
hahahahahaha I got so amused with the bam! and double bam! xD
:)
“Little bam” and “Bam?” 😂
:)
Thanks for the clearest explanation! Like your approach and way to deliver those complicated ideas in a simple way. Please make more videos for all the learners! Thanks!
Thanks, will do!
Your best video I've watched so far. Now I understand perfectly how specificity and sensitivity are obtained
bam! :)
I felt that double BAM in my soul
Awesome!! :)
The theme song today: "wait till u see, ROC" so cool.
Thanks! :)
Spent an hour trying to understand ROC and AUC using other online resources, glad that i finally found this tutorial! You are the best, thanks so much
Thanks!
best ever explanation in the history of Model evaluation and scores! for sure!
Thanks!
Gazillion bam
Hooray! :)
Hello Josh, I am from India. I just started watching your machine learning videos, want to say you are amazing. Thank you for your simple way of explanation.
Thank you very much! :)
Thank you so much for this amazing explanation. This is the first time I actually understood the idea behind AUC and ROC.
Glad it was helpful!
I find myself screaming BAM! when it appears on the screen. Your videos are helping me out so much! Keep up the great work!
Awesome!! :)
Yow I'm just stunned how well you explained these things. Every minute is a euphoria for me. It all makes sense now AAAAAAAAA
Thank you!
I've learned both Statistic and English from this Channel. Double BAMMM!! Thank you all the time for your Awesome video :D
YES! That is a Double BAM! :)
Thank you for explaining this at 11.20pm before my supervision at 9am tomorrow morning! Needed it!
Good luck! :)
I love this channel so much. Every time I start a video, the musical introduction cracks me up
BAM! :)
Josh, you have taken teaching to the next level. Keep explaining and simplifying the complications., it is really really very helpful. Thank you for your contribution in spreading the useful knowledge.
Thank you! :)
@@statquest I always found "confusion matrix" very confusing.... but you solved the confusion 🙂
@@aishah8398 BAM!
These videos helped me BIG time understanding the concepts that are discussed in the text books. These are nice supplemental videos to the text books. I felt easy to understand the text book after/before watching these videos. Ulti BAM 💥
Glad it was helpful!
Brilliant!!! I always come back to this video whenever I need a quick refresher. Thanks so much for your content!! BAM!!
Happy to help!
The was probably the best and most simple explanation I've had, thanks!
Hooray! :)
best tutorial ever watched...now my fundamentals are clear thanks..
Awesome! :)
Thanks a TON! I've never come across such a simple explanation to a complicated problem.
Thank you! :)
I am so happy that this best video got 801k likes, whenever I need to review the content, this one float up to the top as always.
Thank you!
@@statquest YW
That's blowing my mind!!! I'm 31 years old and I finally get it!!! Many thanks to author!!!
Thank you!!! :)