Support Vector Machines Part 1 (of 3): Main Ideas!!!
ฝัง
- เผยแพร่เมื่อ 31 พ.ค. 2024
- Support Vector Machines are one of the most mysterious methods in Machine Learning. This StatQuest sweeps away the mystery to let know how they work.
Part 2: The Polynomial Kernel: • Support Vector Machine...
Part 3: The Radial (RBF) Kernel: • Support Vector Machine...
NOTE: This StatQuest assumes you already know about...
The bias/variance tradeoff: • Machine Learning Funda...
Cross Validation: • Machine Learning Funda...
ALSO NOTE: This StatQuest is based on description of Support Vector Machines, and associated concepts, found on pages 337 to 354 of the Introduction to Statistical Learning in R: faculty.marshall.usc.edu/garet...
I also found this blogpost helpful for understanding the Kernel Trick: blog.statsbot.co/support-vect...
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...
TH-cam 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
0:40 Basic concepts and Maximal Margin Classifiers
4:35 Soft Margins (allowing misclassifications)
6:46 Soft Margin and Support Vector Classifiers
12:23 Intuition behind Support Vector Machines
15:25 The polynomial kernel function
17:30 The radial basis function (RBF) kernel
18:32 The kernel trick
19:31 Summary of concepts
#statquest #SVM
Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/
How much price
Is it available in India
@@duniya-wale Yes, it's available in India at a special price. See: quratebooks.com/shop/book/9798986924021 or www.amazon.in/dp/B0BRQ4K43L
@@statquest How about China?
@@hackerbf8160 It's being translated into Chinese. I don't know when it will be available.
I am a PhD student in the field of data analytics, and took some machine learning classes, but this...
Just WOOWW!
You're a significant value for the world!!!
Thank you very much!!! :)
hangi üni hocam?
The professor makes the most complicated concept the simplest.
M a PhD student in climate change and learnt a lot here... Cannot thank enough!
Uğur Üresin Josh Starmer’s p-value is
Someone said, "You understand a subject when you can explain it to a 3 year old". Your teachings show how well you have understood the things you teach us! Great job!!
Thank you very much! :)
I think it was Richard Feynman
That's is true! The people like Mr. Starmer must be awarded by gold for such an awesome teaching. Supertalented professor!
@@AlvaroALorite einstein also said something similar but that had 6 year old.
The teachings are simply perfect..I recommend Statquest to everyone 👌
This might be the greatest source of knowledge on the internet. God double bam bless you Josh
Thanks! :)
@@statquest But you only add more dimensions when you have more than 2 categories like sick or not sick or more than 2 propertie being measured right?
@@leif1075 The number of dimensions doesn't depend on the number of categories or properties being measured. In fact, the RBF has infinite dimensions. For more details, see: th-cam.com/video/Toet3EiSFcM/w-d-xo.html and th-cam.com/video/Qc5IyLW_hns/w-d-xo.html
@@statquest then why the squaring then..thought it wasonly when there was 2 properties like you said in the video?
@@leif1075 What time point in the video, minutes and seconds, are you referring to?
Beautiful work Josh. You have an extremely rare gift of opening up the black box that few people are brave enough to do, much less try to present it in plain English. I especially appreciate how you don't dump a bunch of summations and Greek symbols on learners. I may have to contribute more support after this video.
Thanks so much! I'm glad you like this new one.
Pandora box? no it is StatQuest box once u open it BAM!
I'm SO glad I found this video! The clear explanations make it easier for me to pick up the concepts. ^_^
My lecture notes are filled w the greek symbols, and I don’t understand shit in those.
@@aarnavgupta3936 one of my most favorite, honest statements ever lol
Andrew Ng got me off to a good start a few years ago, but Josh is hitting it home for me. Thank you!
I'm glad the video is helpful! :)
I have trouble with understanding Andrew Ng. Do you think watching Josh's videos will help me understand Andrew's better? I am trying to get a little deeper into the mathematics.
@@yancheeyang2918 Depends on where you have to start, I would suggest Khan Academy and doing their mastery exercises (You need Algebra 1 and 2, statistics and probability calculus and multivariate calculus) , and for a linear algebra course I suggest the MIT open library course.
On Andrew course you use octave but I suggest you take MIT intro to programming with python.
@@isaacpicon169 Andrew confuses the shit out of me, especially in neural network and SVM
STATQUEST STOP showing up when I search up complex topics that my lecturer cant explain IM TIRED OF YOU HELPING ME OVER AND OVER AGAIN
Every word my lecturer says is a terminology alert
bam! :)
"BAM" is a perfect way to let someone know they just missed some key information. Great video
Thank you!!! :)
This is literally the first time I'm actually commenting on TH-cam. You, sir, have achieved, single handedly, something that my professors, textbooks and several additional resources have failed, in 20 minutes. The level of understanding that you have of Machine Learning is just remarkable, and your delivery makes sure I don't get bored, no matter how long the study session is. Can't remember opening Part 2 of an educational video with such enthusiasm. Will support as soon as I can. Keep it coming!
Thank you very much! :)
ULTIMATE BAM: You make me understand kernel trick in few minutes where nobody does !
Hooray!!! Thank you very much! :)
@@statquest where can I find your kernel trick video?
Meta-Bam to the infinite nth degree.
@@lashlarue7924 Hooray! :)
I really like how Josh breaks the concept into smaller, easily-understandable parts. All the 'psst' and 'bams' may sound disruptive, but they actually worked well by signalling the important concepts the viewer should know. Good job!
Thank you very much! :)
It's annoying and cringey.
this bam and pist is important some people. i like those sound. its not disturbing me. i am enjoying it.
After watching a number of tutorials explaining SVMs, this by far is the best one.Thank You.
Thank you very much! :)
At this point, I have come to agree that whenever I'm searching for a video to better understand any concept and I see Josh Starmer's in the search result I become confident I will understand that concept. Amazing!
Wow, thanks!
I consider myself a competent programmer and reasonably good at math, yet I couldn't grasp how SVMs work only by reading my textbooks. This video is so much better than the several pages of formal definitions and mathematical proofs -- formality is very important but it's not the best way to introduce new concepts for learners.
You did an amazing job and helped me a lot. I'll be buying one of your t-shirts as soon as I get some money. Thank you!
Thanks!
This 20 minutes presentation give me more clarification than my 3 days effort on other sources. Extremely grateful.
Awesome! :)
Absolutely spot on.😂
@@mujeebrahman5282 Thanks! :)
You're the best, Josh.
Your teaching methodology should be the standard for any undergration courses.
Please, keep doing this amazing job.
Wow, thanks!
I'm so glad to finally find the channel about ML where the speaker actually understands the topic and thus able to explain it in the concise but well-explained manner. It also helps tremendously that it's by somebody whose English I understand. Most of videos I've found are lacking in both.
Thanks!
Hello, I think the least I can do is to thank you for all your videos:
1) You make complex stuff just so easy
2) You explain very clearly and provide nice and well-explained examples in which you show how proficient you are in all these things
3) Your little jokes make us enjoy the lessons so much
4) You save us all from our tricky situations like complex assignments without having a proper explanation and feeling as if we could never understand this stuff.
Thanks again in case I have not expressed my joy clearly enough
Hooray!!! I'm glad you like my videos so much! :)
This is beyond my imagination. You're blessed to teach and make people understand really really well without any human interactions. Keep it up! God Bless You! Cheers!
Thank you very much.
More about kernels plz.this one is also a gem from u
I was scratching my head for the past few days on SVM and almost lost hope. Man I just found your channel. Got cleared right away on the first go. Thanks a ton. God Bless.
Glad I could help!
You're a genious man. In this field where most reports are too complex and tiring to read, this video actually gave me more energy while learning. Nobel prize for you
Awesome! :)
This is easily the best SVM explanation. It's far better than the lecture at MIT about it I took years ago, which says something. MIT often has some of the better lectures out there, but this blows it away. Thanks Josh S, keep up the great work!
Thank you very much!!! :)
I notice you interact with every comment - I just want to say the attention you put to your videos and the community you've built is amazing, you've made me love data science and I wouldn't have given it a chance if it wasn't for your explanations
Thank you very much!!! I'm glad my videos are helpful. :)
I didn't understand my professor's lecture entirely so I decided to venture onto TH-cam. I am so glad that I found your channel! I actually UNDERSTAND it!!! You make STAT so much easier to understand!!!!
Happy to help!
You are a true savior, almost all tutorials and courses begins with a bunch of matrices and Greek letters, which is extremely painful to me. I love how you start with 1 dimensional graphical example, which is super straightforward, and your teaching method inspire viewers to actively think instead of feeding us stuff. The visuals are super neat, simplistic and well designed, and I love the humorous language you use to actually make math interesting. Awesome job!!
Thank you very much! :)
BAM!!! This is the most intuitive SVM tutorial I've ever seen. Thank you for your contribution to humanity.
Thank you! :)
When you see that learning could be so enlightening (and entertaining) your motivation skyrockets. Thank you!
Thank you very much! :)
This is just amazing. As they say, if you can not explain something in easy way, you don't quite get it... I wish at least 10% of school/university professors had your gift!
Thank you! :)
That was a very clear explanation, The best part of the video is when he says "... I don't know how to draw a 4-Dimensional graph..." :D
bam! :)
my Intro to ML was just ridiculous and this thing is GOLD. Pretty much all the info gets injected to my veins directly so I don't think I'm forgettjng any of this any time soon!
Hooray!! That is awesome. I'm glad you like the video. :)
After reading through SVM algorithms for the majority of the day not knowing what was going on exactly, this has been extremely helpful. Thank you!
Glad it helped!
The most clear explanation of SVC and SVM with simpler jargons I have ever seen. Thank you
Thanks!
Oh my gosh! I opened up youtube and here is StatQuest clearly explaining SVMs, could my day get any better? I've been waiting for this! thanks a lot, josh.
Hooray!!! :)
Finally, you're explaining SVMs! Been waiting on this for a while
Thanks!
For the first time, I am beginning to understand AI/ML, thanks to your videos.
Happy to help!
What a nice approach. A lot of books in a 20 minutes video, not missing ANY important info. You are an artist. Thank you.
Wow, thank you!
Wow, I'm trying to learn my whole "Machine Learning" lecture in one week, because I hadn't much time for it this year due to other lectures which killed my time. You are saving the exam I have next tuesday, this video is MEGA BAM! I will check out more of your videos after finishing this comment.
It must have been a big effort making this, thank you sooo much! I will provocatively send this to my professor and ask him, why he can't explain it like this. BAM!
BAM! :)
I am watching it for the n-th time and everytime I tend to understand the concept better and better. Awesome Josh!
Hooray! :)
This is probably the most elegant explanation I have seen on SVMs. I feel like I have hit a jackpot after coming across this channel. You sir, are a LEGEND! BAM!!!
Wow, thanks!
Been following your videos for some time now. You truly are gifted to be able to simplify these advanced concepts for viewers like me who tend to get stuck trying to learn it mathematically at first. Absolutely brilliant content, with a manageable pace, great visuals, clear delivery and the slight humor added. Quadruple BAM!!!
Hooray!!! Thank you very much! :)
Amazing sir, I am currently pursuing an MBA in Finance(Major) and Analytics(Minor), but I have never seen such types of explanation. Love you from India Sir.......
Thank you! :)
This is truly the best explanation of SVM I've ever seen! Thank you very much!
Hooray!!! :)
Every time I'm about to give up on my quest to learn machine learning, you come to the rescue!!! Thank you so much!!
Hooray! :)
I hold a Robotics Ph.D. degree from a prestigious university and I have to say that this video is way better than the lecture I took in my graduate school. Good job!
Thanks!
Hi Josh, thank you for making these amazingly videos. I took a machine learning class where the professor made slides from the Pattern Recognition and Machine Learning book by Christopher Bishop. I could not understand the material very well back then. But you have made it much clearer by using visual examples to give an intuition behind what the math.
Thank you, and thank you for your support! :)
I was sitting there desperately trying to revise and cram in all content before my postgrad Statistical Machine Learning exam. I then heard a scratch on my window, I was in disbelief upon realising it was Josh! In my awe-struck state I fiddled about and managed to open the window, Josh climbed through. "How did you get to Australia so quickly?" I shakenly asked. "Non-linear models" Josh answered, of course it was non-linear models, how could I forget such a sim- "Don't be embarrassed, we're all noble adventures on the Quest of Stat" Josh kindly said. I was nervous at this point, I didn't really know what to do, but I felt Josh's presence to be a benevolent one. Josh then approached and whispered in my ear "double bam", and we were then immediately sent to the astral-hyperplane in which we skipped around in checking out representations of models in 4d, 5d and even -3d realms.. I cannot describe such a place, it was beyond human comprehension. I still question whether or not I was meant to see such a realm, but regardless I have faith in Josh. After what felt like an eternity, Josh told me it was time to return to the 3d, regular plane realm, well he reminded me that it's still a hyperplane but because it's ordinary to not worry about that.. but that's not the point. I asked if I would pass my exam and Josh said that he classifies it with a high likely-hood. I also asked if anyone would believe my story and Josh said "no bam". I then woke up at my desk, and 5 minutes had passed since Josh supposedly arrived at my window. I guess I need more sleep.
By the way, great video. You're a legend Josh!
This is the best comment I have ever read. You are true poet!!! :)
Well that's very kind. :) I hope my sleep deprived comment gave you a good laugh. And thank you for all your hard work! It's clear everyone, including myself, is very grateful and appreciative!
My only hope is that you pursue writing at a serious level! It brought a lot of joy to me.
This man has beautifully submerged ease and lightness to dissolve some unconsciously built weight and stress of concepts.
:)
I always play videos in 1.5x but I played every video I watched of yours in normal mode because the way you taught was brilliant and it got me held and I didn't move till I finished it. Great job Josh. You made machine learning my favorite because of your way of teaching. God bless you. Keep posting more content :)
Awesome, thank you! :)
🙏🏻Great. I had been postponing reading on SVM...then finally started with this video. It explains things that would have Einstein’s approval ‘if you can’t explain it to a 5 year old, you haven’t understood it’.
This video can be understood by anyone and everyone.👍
Thank you very much! :)
Wow, this was my first video from StatQuest and I have never done this before. But i subscribed immediately after watching the video. Excellent explanation !
BAM! :)
I have exams tomorrow revising through your videos made it 10 times easier thanks man
Best of luck!
You are a gift to the students struggling to understand these concepts which universities are failed to deliver!
Thank you!
woah! i finally understand what i've been trying to learn for the past week in my ML class. your channel is a godsend
Thank you very much! :)
The BAM keyword makes more focused to video!
[ Terminology Alert : please don't read comments while watching video ]
Nice! :)
srsly but overdose of bams make me obsessed with it but a little ........ too.
Amazing video ! The fact that everytime you go from low to high dimension to keep it understandable is reaaaally nice.
Thank you! :)
OMG...!! how can someone teach a complex topic in so easy way...U r doing a great job....I searched n no of tutorials and finally got StatQuest ...and then i never search for any other channel...Thanks alot Josh..You are a Gem
Thank you!
Clap, clap, clap! I loved your explanation. Fantastic! Thank you very much!
Thank you! 😃
I have no words other than: WOW, Bam! Thank you! What you are doing is amazing!
Thank you very much! :)
I'm speechless! This video, you, and your humour! You're a gift to the nerdy/solid world of CS! Thank you very much for your time and effort on videos like this. It was simply amazing. AMAZING!
Wow, thank you!
Hi Josh - I would like to say a huge "Thank You" for what you do. These videos take an amazing amount of work and expertise, and I really can't think of any other medium that comes close. Even when the math is above my level, I can follow "small bam" / "Bam" / "BAM!" and pick up the nuances from there.
So in short - thank you for what you do, your material could cover 4 years of college 5-times over, and I appreciate the monumental effort you put into these videos. THANK YOU!
Thank you very much! I'm glad you appreciate all the time I spend on these videos. You are correct - each one takes forever to make - but I love doing it. It's my favorite thing to do.
This is the best explanation of SVM .
I'm not just talking about this single video when I'm saying thank you for helping me pass my Machine Learning course 🥰 Btw, I bought and absolutely loved your Jupiter notebooks and study guides 🙂
Wow!! Thank you very much and thank you for your support!
This is one of the most lucidly explained machine learning channels. I am a big fan..... god bless you josh....
Thank you! :)
You, sir, deserve respect. I read so many resources. attended so many lectures to no avail.. You are a master at your job
Wow, thanks!
Wish more people talked to me like Im a five year old! Really great video, thanks!
Incredible job. Way to simplify ML concepts to the point where a 10 year old can understand them.
Thanks! :)
You saved my day. You are amazing. I literally have no words to thank you. In my journey into the amazing world of data science, your channel is the best resource I could ever find. I actually came here to understand SVM while learning the ML Course from Stanford on Coursera. I could not understand all the long mathematical equations there but now, I feel very confident. Thank you so much for creating this content and sharing it with us. God bless you always.
Thank you very much! :)
Hands down you are the greatest educator I have ever come across. Amazing work, thank you!
Thank you very much! :)
Just wanted to state that a 'like' and comment here wont convey how thankful i am for this video :D Keep up sir !
Thanks! :)
BAM!!!! I was waiting for this video for so long. Thankyou Josh!!!❤️ Love from India❤️
Hooray!!! :)
Your in-depth understanding of the concepts is the real reason behind such wonderful videos Josh.
Thank you!
knowledge is not just read a lot of things and memoise it. you have shown what is knowledge. the way you made this video is just wow. it was amazing when the high echo sound of support vercor machine😂. its like first entry of a hero in a film
Thank you!
Would appreciate a video on word2vec, and thanks for the awesome content 😃
Absolutely clear and magnificently explained, I feel like the word "Kernels" should be added to the title of this tutorial :), Double Bam !!!. That was great Josh, keep it up
Thank you very much! I have two follow up videos that dive into the details of the polynomial kernel and radial kernel if you are interested.
@@statquest I can't wait to watch them :)
@@MohammedBakheet Here's the link to the polynomial kernel: th-cam.com/video/Toet3EiSFcM/w-d-xo.html ...and here's the link to the Radial kernel: th-cam.com/video/Qc5IyLW_hns/w-d-xo.html NOTE: They are a little math heavy, so let me know if you have trouble with them.
It's amazing you got 600k people to watch a video on this topic. I didn't expect there to be 600k people on earth that understood this topic.
:)
In just 20 min, I clearly understand what SVM is, thanks to your wonderful presentation. You are a awesome teacher!
Thank you very much! :)
That was the perfect "Helloooo" to get anybody started off with a study session!!!!!
Bam!
This video was incredible. Thank you so much for that !.
Best Wishes and health in these quarantinetimes.
Marcel.
Thank you very much! :)
Soo grateful for the work you do!!! Working on my masters in analytics and you consistently distill the topics I find myself struggling with in lectures into the simplest form - thank you
Thank you very much!
I'm a big fan of your intro music. I think it adds character and it seems not many people have that courage.
Thanks!
This IS AMAZING MAN! Keep up the great work!!
Thanks! :)
14:18 BAM indeed! a really beautiful, mind-blowing scene for first-time learner like me!
Hooray! Thank you very much.
From what I have seen, you have the most elegant explaining skills of all. You are a BAM! to all the people out there who are trying to look smart or something by just confusing us newbies. I actually found your video link in the comments section of a paid course(A small bam may be?). Great job Josh!
Thank you very much! :)
Omg I can't thank you enough. I couldn't wrap my head around SVM no matter how hard I tried, the math and the terminology always overwhelmed me. But your videos are so easy plus you make it fun with your music and jokes and the special effects- Support Vector Machines (echoes)
Thanks for such amazing videos. I wish you all the best in life. You are an angel for ML enthusiasts
Thank you very much!
I'm in love with ur way of explaining 😍
Hooray!!! Thank you very much. :)
I need a refund on all my math/stat/prob courses. I found a gold pot.
:)
wow what an absolute goldmine of a channel this is
Thanks! :)
one of the best explanation of Support Vector Machine that i've read, so visual, so easy to understand in very short amount of time. You are my best hero now, will checkout every single videos in your channel, thanks a lot Josh, keep posting.
Thank you very much! :)
Great video man, it was very clear and easy to keep up. Thank you!
Thank you!
Hello sir
I belong to India
I like your videos 😊 because you explain everything clearly
You reduce tough things into simple form
Sir also make videos on (Negative, geometric and hypergeometric) distributions only if you have free time
These topics are not available in your playlist
I'm glad you like my videos! Thank you. :)
What an amazing video. Thanks a ton man. As a software engineer with some small knowledge of machine learning in the back of my head, this is incredibly useful. Exactly the right speed, nice pictures and great narrating voice. Also the little jokes now and then help A LOT for a nervous ADHD kid like me. Thanks man.
Awesome! Thank you very much! :)
I'm currently taking a master's level text mining class. This just made the concept more clear than anything I've seen before. Thank you!
Thanks! :)
I can tell that is the most passionate 'Yesssss' in the world
YES! :)
Psst! In TH-cam jargon, That's the best video about SVM
Thank you so much! :)
excellent video, no use trying to read a book with hundreds of pages when someone as brilliant as you did all that in a matter of minutes. you will be seeing a donation as i finish my thesis. i kept it as a note on my noticeboard for future reference. thanks again
Wow, thank you!
This is the clearest explanation available in the world
Thank you!
When my Machine Learning professor is being a total noob, I come to the StateQuest for help.
Awesome!! :)
3:30 - "bam? No. No bam." 😂😂😂
:)
I just love your sense of humor and additionally the simple but very intuitive explanation you have for those things! very recommendable :)
Thank you very much! :)
I really love how you intricately explain extremely complex ideas in great details like they elementary knowledge (SVMs are only taught at university, often at the grad-level)!
Thanks!