Judging by how much you simplified the lesson without sacrificing the actual concept, I trust you have spent good time preparing for your lesson and animation. Thank you very much
I haven't seen a better explanation for PCA. Love it! And I always had trouble to conceptually understand covariance. I knew it was related to direction but always found it difficult to conceptualize. Thank you for your explanation.
we may be aware of terms like mean, variance, covariance, linear-transforms, eigenvectors and eigenvalues, PCA. This video puts together all the relationships in an amazing way
I am NOT watching his video because I wanted to learn math or Machine learning. I am already pretty good at math and machine learning. I am watching these Luis Serrano videos to learn how to effectively teach this stuff to young people. Those who might think that they can't do machine learning or math. I sent his book to my son, a business/communication major in university, he is now studying machine learning with the help of Luis Serrano's machine learning book. Thanks
I am impresed with the level of information you were able to put together in a 20 minutes video. I mean you summarized a whole semester in such a way that a beginner could follow. Thats amazing
This is the clearest explanation of PCA I've ever seen. As a wannabe analytical chemist, this is pure gold because it gives you hints about the true meaning of PCA without excessively digging into statistics and pure mathematics which just blows your mind. Congrats for the amazing explanation and the effort to make this video!
PCA, that I couldn’t really understand in five years (although I have been using in analysis without being clear), I am now crystal clear. There are so many articles and videos. Those are like a parroting. This lesson is the best.
Great explanation Luis! It took me a whole class and a conversation after with an instructor to get PCA, and still your video cleared up so much uncertainty/misconceptions I had on the subject.
Most comprehensive explanation, thank you so much. Tomorrow, I have interview for big tech company and now after watching whole video, I feel confident to answer questions on PCA. \m/
This is my first comment ever in youtube ,I was searching for PCA videos and was unable to figure out how exactly PCA works,but this video gave me complete clarity on the topic....Thanks Luis Serrano!!!
This was AMAZING!!! You reconciled so many disparate concepts I had learned in college...I understand how it fits together now. Very down-to-earth explanation, and amazing animations!
HIT THE LIKE BUTTON IF YOU LOVE THIS VIDEO!!! You did it! You nailed it!! You killed it!!! Just when I was about to give up on this topic after watching a series on PCA, I ran into yours and BOOM!!! IT WAS THE BEST THING I DID!!!! I am going to watch it again because you gave me an ideal explanation of the topic. You started off with related background knowledge necessary for perfect understanding. THANKS A LOT!
I have spent 3 days chasing eigenvectors and eigenvalues, and then I found you video. It puts it all together very nice, and mentally delivers it with ease - like a boss. I was very close to spending two weeks on Khan academy to understand what you delivered in 26 minutes. Thank you so much.
This excellent, Luis. My only thought to improve the video would be to pull the thread through using your original example of housing data. Good job, and thank you!
This is great tutorial. Can you make a similar video about - Markov Chain Monte Carlo, since its too complicated to understand why are we doing what we do?
Luis, ONE OF THE BEST PCA EXPLAINED ON UA-cam ARENA OF MACHINE LEARNING. I read many blogs, almost all the PCA tutorials on youtube but I was unclear about the PCA concept, was thinking what is this PCA.. but when I started viewing your PCA explanation , within 20 seconds PCA lighten up in my mind, which started from the taking picture example. The way you explained this confusing concept, It's totally insane, You are an exponentially amazing tutor. many thanks for sharing this stuff. you are simply amazing :) Thank you so much Luis.
Recognizing your name from your Udacity lectures, I was thrilled to have stumbled upon your youtube channel today - really gentle, concise, intuitive explanation of a complex topic. Nice shout out to 3blue1brown - watching this reminded me of his videos and style (though his visualizations are insane :] ). One question: from the end of your video, it seems to me that eigenvalues can be thought of as measures of feature importance, but in your housing example, multiple similar features (size, #rooms, etc) become a new single representative feature. How do we know what the features are that we are left with after PCA in much higher dimensional examples? Say we reduce a 100 dimensional data set to 10 dimensions, how do we know what those 10 dimensions represent? Thanks again - looking forward to diving into your channel.
footfalcon thank you for the kind message! (And sorry for the late reply). The new features are combinations of previous ones, but in general they may not mean much. However, since you know what combinations they are, you can look for example at the biggest coefficients in the linear combination, and see if you find trends. For example, if the top feature has big coefficients for several features that are size related (in a house) and small coefficients for other features that are not related to size, you can associate that feature with size in general. But it’s more of an observation than a mathematical way to do it.
I've watched so many videos, read through so many websites, and asked so many questions to chatgpt about pca. I have to learn it as quick as possible for a research program. This is by far the best explanation I've seen!!
Thanks I Spent lakhs for an Analytics course and was still at loss for what is Eigen Value and Vectors. You made it so clear. Even instructors having Stats majors and years of teaching experience don't know their stuff or lack the discourse.
This is one of the best tutorials which I have ever seen. Being a teacher, I appreciate the way how difficult concepts are explained in a simple way. It is must-to-read video for all students who want to understand PCA.
Best explanation of PCA till now i saw many videos on pca i always look at time when they will finish because they are not clear but your video im not bother about time because it is simple n clear, i will watch all your videos.Thanks alot Luis
oh my god! this is unbelievable stuff. I wasted around a day on many lectures and books still can't able to grasp a shit about PCA. but this man explains it in such a beautiful way. You're a legend sir
I finished watching this video at 3am, not because I have any deadlines to take care of, but simply because I wanted to gain deeper understanding on PCA out of my own desire. And your video was simply the best. No, really, the best. It has just the right amount of math, it builds up all the small steps, it capture the essence of all those small steps, etc. Just a perfect balance between theory and plain (English) language. Your explanations are fantastic and your tone is just so perfectly calm. Keep up the good work!
This is the best explanation of PCA, individual components broken down, emphasis on visualization instead of just the method and then all the components put together for the big picture. I cant thank you enough for the fantastic teaching style that clicks with me.
This is how you teach!! I wish my professors from university could do even 1% of this... I am paying thousands of dollars for something I get for free on youtube!! Very good job!!
I spent so many hours looking at other videos and content online and could not make sense of it. This video really helped me a lot. Especially in understanding the role of eigenvectors and eigenvalues. Thank you so much!
This video is literally a life saver for me! I'm learning Applied Machine Learning and Data Mining and have seen couple of videos on PCA and was more and more puzzled after watching them. Believe me I have not found any video better than this. You've provided the detailed explanation on all the terms like Variance, Covariance, Linear Transformation, Eigenvalues, Eigenvector and able to relate them. In more than 3-months, my professor was not able to show the relation what you've shown in 26 minutes straight! I wish I would have seen this in my previous semester to understand why we are learning Linear Transformation and Eigenvalues and Eigenvector in math syllabus. I would suggest to put in the description the practical use of Linear Transformation, Eignenvalues, Eigenvector so that students like me can relate this. Thank you once again!
By far the best basic explanation of PCA on the internet! Thank you so much, I did not think I could ever understand this without being an expert mathematician.
All these UA-cam videos are great. There is a 100% mapping between this and all the AI stuff with real brain observations and computational neuroscience. This author is brilliant!
the best explanation I've ever come across, either in statistics books or on the Internet. You have extraordinary didactic skills. Thank you very much.
The skeptic in me wants to say I understood most clearly from this lecture because I had a solid subconscious understanding from watching probably a hundred of these over the years The other part of me tells the skeptic to shut up and appreciate the magnificence of this presentation. This is a great example of data storytelling from my view.
Data Science program + other PCA videos != Understanding of PCA. This explanation was amazingly clear and concise without feeling like I'm being coddled. Just subscribed!
I hated eigenvectors and eigenvalues for 20 years because I thought they were useless. Now, I have an appreciation for these concepts because of your video. Thank you!
Masha Allah, sir. What a presentation! I have not encountered someone who could explain this topic in such a way that even a novice like me can grasp it.
I was loitering here and there from one UA-cam channel to another to get a better explanation of algorithms, and finally I landed up here...sir, your explanations are so easy to understand,thank you for your efforts ❤️
Thank you very much! I hardly comment on UA-cam but when you see real skills, you have to acknowledge. I have watched several other videos on PCA but didn't really get what was going on. This video is so simple and clear for anyone with elementary knowledge in geometry and statistics to understand. Thank you very much !
I haven't seen any clearer explanation of PCA on the internet than this one. Many thanks!
I agree!
I second this - great channel!
agreeeeeeeeee
Completely Agree!
+1
Judging by how much you simplified the lesson without sacrificing the actual concept, I trust you have spent good time preparing for your lesson and animation. Thank you very much
I haven't seen a better explanation for PCA. Love it! And I always had trouble to conceptually understand covariance. I knew it was related to direction but always found it difficult to conceptualize. Thank you for your explanation.
@@rishabhkapoor3338 www.amazon.com/Data-Analysis-functional-principal-regression/dp/B088BM4FCB/ref=sr_1_6?dchild=1&keywords=nizar+soilihi&qid=1589823715&sr=8-6
Excellent video. Very intuitive. Thanks
hi
I have confusion even after watching videos of 2+ hours...
26 minutes solved all doubts.
I like the way how you put all the concepts together.
Thanks.
lol this cracks me up
we may be aware of terms like mean, variance, covariance, linear-transforms, eigenvectors and eigenvalues, PCA. This video puts together all the relationships in an amazing way
I am NOT watching his video because I wanted to learn math or Machine learning. I am already pretty good at math and machine learning. I am watching these Luis Serrano videos to learn how to effectively teach this stuff to young people. Those who might think that they can't do machine learning or math. I sent his book to my son, a business/communication major in university, he is now studying machine learning with the help of Luis Serrano's machine learning book. Thanks
You are the tutor we all need but don't deserve.
Best video i have even seen about PCA. Thank you so much sir.
I am impresed with the level of information you were able to put together in a 20 minutes video. I mean you summarized a whole semester in such a way that a beginner could follow. Thats amazing
i search all the pca tutorial on youtube, and this is the best one.
This is the clearest explanation of PCA I've ever seen. As a wannabe analytical chemist, this is pure gold because it gives you hints about the true meaning of PCA without excessively digging into statistics and pure mathematics which just blows your mind.
Congrats for the amazing explanation and the effort to make this video!
I am not sure why this video is not appearing in top search results under 'PCA'. Best video. Period.
That arguably might be the best tutorial I've ever watched on UA-cam. Thanks.
PCA, that I couldn’t really understand in five years (although I have been using in analysis without being clear), I am now crystal clear. There are so many articles and videos. Those are like a parroting. This lesson is the best.
Great explanation Luis! It took me a whole class and a conversation after with an instructor to get PCA, and still your video cleared up so much uncertainty/misconceptions I had on the subject.
Thanks! Very good explanation of how projections, eigenvectors and eigenvalues all come together in PCA analysis!
far better PCA explanation on youtube, I saw a lot of videos, but only learned with this one.
This is without a doubt the best explanation on youtube.
Most comprehensive explanation, thank you so much. Tomorrow, I have interview for big tech company and now after watching whole video, I feel confident to answer questions on PCA. \m/
Thank you! I hope your interview went great!
This is my first comment ever in youtube ,I was searching for PCA videos and was unable to figure out how exactly PCA works,but this video gave me complete clarity on the topic....Thanks Luis Serrano!!!
This was AMAZING!!! You reconciled so many disparate concepts I had learned in college...I understand how it fits together now. Very down-to-earth explanation, and amazing animations!
Only explanation which has core PCA explained in a very simple yet effective and useful manner.
His teaching skills are exemplary.
This is the best PCA video ever seen, it's quite easy to follow and understand deeply.Thanks!
This is so far the best video on PCA.
The first 20 mins is like sharpening your axe and last 6 mins- cut the tree.
Awesome explanation for PCA beats any in youtube tilll date
EXACTLY!!!!
Agree, even the one from 3Blue1Brown
most simplified explanation of PCA👌👍
HIT THE LIKE BUTTON IF YOU LOVE THIS VIDEO!!!
You did it! You nailed it!! You killed it!!!
Just when I was about to give up on this topic after watching a series on PCA, I ran into yours and BOOM!!! IT WAS THE BEST THING I DID!!!!
I am going to watch it again because you gave me an ideal explanation of the topic.
You started off with related background knowledge necessary for perfect understanding.
THANKS A LOT!
My goodness, the video was like a movie. All things in correct order to understand PCA. So well explained. Thank you so much, keep up the great work!
well thats the best explanation i have ever seen, great job sir, thank you so much !
I have spent 3 days chasing eigenvectors and eigenvalues, and then I found you video. It puts it all together very nice, and mentally delivers it with ease - like a boss. I was very close to spending two weeks on Khan academy to understand what you delivered in 26 minutes. Thank you so much.
wow! this was VERY worthwhile. thank you! so many lights switched on
Deepest gratitude from someone who’s just begun learning ML 5 years after this video was posted 😂 Thank you sir you are a lifesaver
This excellent, Luis. My only thought to improve the video would be to pull the thread through using your original example of housing data. Good job, and thank you!
After watching 3Blue and 1Brown, this video made the most sense to me. And I had watched many videos on PCA. This is the best one
This is great tutorial.
Can you make a similar video about - Markov Chain Monte Carlo, since its too complicated to understand why are we doing what we do?
f
Same here, really would love a video on this topic
up
I second that
Luis, ONE OF THE BEST PCA EXPLAINED ON UA-cam ARENA OF MACHINE LEARNING. I read many blogs, almost all the PCA tutorials on youtube but I was unclear about the PCA concept, was thinking what is this PCA.. but when I started viewing your PCA explanation , within 20 seconds PCA lighten up in my mind, which started from the taking picture example.
The way you explained this confusing concept, It's totally insane, You are an exponentially amazing tutor. many thanks for sharing this stuff.
you are simply amazing :) Thank you so much Luis.
Recognizing your name from your Udacity lectures, I was thrilled to have stumbled upon your youtube channel today - really gentle, concise, intuitive explanation of a complex topic. Nice shout out to 3blue1brown - watching this reminded me of his videos and style (though his visualizations are insane :] ). One question: from the end of your video, it seems to me that eigenvalues can be thought of as measures of feature importance, but in your housing example, multiple similar features (size, #rooms, etc) become a new single representative feature. How do we know what the features are that we are left with after PCA in much higher dimensional examples? Say we reduce a 100 dimensional data set to 10 dimensions, how do we know what those 10 dimensions represent? Thanks again - looking forward to diving into your channel.
footfalcon thank you for the kind message! (And sorry for the late reply). The new features are combinations of previous ones, but in general they may not mean much. However, since you know what combinations they are, you can look for example at the biggest coefficients in the linear combination, and see if you find trends. For example, if the top feature has big coefficients for several features that are size related (in a house) and small coefficients for other features that are not related to size, you can associate that feature with size in general. But it’s more of an observation than a mathematical way to do it.
I would literally clap for you if you gave this lecture in person. Thanks for this man. Keep doing what u do
Thank you for taking time to explain small detail
I've watched so many videos, read through so many websites, and asked so many questions to chatgpt about pca. I have to learn it as quick as possible for a research program. This is by far the best explanation I've seen!!
Thank you!! Really simple explaination of such complex topic.
Thanks I Spent lakhs for an Analytics course and was still at loss for what is Eigen Value and Vectors. You made it so clear. Even instructors having Stats majors and years of teaching experience don't know their stuff or lack the discourse.
You have a very intuitive way of teaching things that's hard to infer when reading or watching lectures from other teachers.
This is one of the best tutorials which I have ever seen. Being a teacher, I appreciate the way how difficult concepts are explained in a simple way. It is must-to-read video for all students who want to understand PCA.
Best explanation of PCA till now i saw many videos on pca i always look at time when they will finish because they are not clear but your video im not bother about time because it is simple n clear, i will watch all your videos.Thanks alot Luis
oh my god! this is unbelievable stuff. I wasted around a day on many lectures and books still can't able to grasp a shit about PCA. but this man explains it in such a beautiful way. You're a legend sir
awesome. it was the best video in definition of eigenvalue and eigenvector. Thanks.
that explanation man .. just wow .. Thank you so much.
The most informative but graspable explanation of PCA I have encountered online ever. Thanks so much!
Quite literally the best video on PCA. You explained everything without skipping the math and the basics, amazing!
Fabuolous video and clear explanation. You made PCA simple and fun
This is the best PCA explanation. If only I had this explanation 11 years ago...
I finished watching this video at 3am, not because I have any deadlines to take care of, but simply because I wanted to gain deeper understanding on PCA out of my own desire. And your video was simply the best. No, really, the best.
It has just the right amount of math, it builds up all the small steps, it capture the essence of all those small steps, etc. Just a perfect balance between theory and plain (English) language. Your explanations are fantastic and your tone is just so perfectly calm.
Keep up the good work!
I am a teacher. And I tell you this. You are the best!!! The absolute best over all!
This is the best explanation of PCA, individual components broken down, emphasis on visualization instead of just the method and then all the components put together for the big picture. I cant thank you enough for the fantastic teaching style that clicks with me.
This is how you teach!! I wish my professors from university could do even 1% of this... I am paying thousands of dollars for something I get for free on youtube!! Very good job!!
My favourite explanation of this topic in the whole of the internet. And I have seen A LOT!!!
Out of the world explanation for PCA. Never seen so simple and easy elaboration of this concept. Tons of thanks.
I spent so many hours looking at other videos and content online and could not make sense of it. This video really helped me a lot. Especially in understanding the role of eigenvectors and eigenvalues. Thank you so much!
The best explanation for PCA. I can't remember concentrating that much for 26 minutes in any of youtube videos. ❤
I haven't seen any better explanation for PCA. This was Brilliant. Hats off for such a clear explanation of complicated topic.
have never seen a better explanation of pca than this
This was simply a great explanation of a complex topic. Thanks!
Probably the best explanation of PCA. Mean, variance, covariance, correlation such a smooth explanation, great.
Thank you . Thank you. Thank you. Really helped. Was stuck in this topic.
Many thanks for this amazing explanation .. Good bless you
Wow..One of the best explanation about PCA..The best part is that you explained clearly the purpose of PCA in simple way!!!Billion thanks.
This video is literally a life saver for me! I'm learning Applied Machine Learning and Data Mining and have seen couple of videos on PCA and was more and more puzzled after watching them. Believe me I have not found any video better than this. You've provided the detailed explanation on all the terms like Variance, Covariance, Linear Transformation, Eigenvalues, Eigenvector and able to relate them. In more than 3-months, my professor was not able to show the relation what you've shown in 26 minutes straight!
I wish I would have seen this in my previous semester to understand why we are learning Linear Transformation and Eigenvalues and Eigenvector in math syllabus. I would suggest to put in the description the practical use of Linear Transformation, Eignenvalues, Eigenvector so that students like me can relate this.
Thank you once again!
You explain the concepts sooooo simply. Hats off! Funny how I saw this video and then took your course on Coursera and it led back to this video .
By far the best basic explanation of PCA on the internet! Thank you so much, I did not think I could ever understand this without being an expert mathematician.
This has to be the best, yet so simple to understand explanation of PCA I have ever come across!
Very well explained in simple terms
I needed to understand how pca is used to calculate the normals of a point cloud and by far this is the best explanation I've seen. Thank you!
You are a great teacher, teaching is an art and this guy is an artist. Million Thanks Sir.
This is the best explanation for PCA. I have tried with so many videos, but this saves me. Thank you
All these UA-cam videos are great. There is a 100% mapping between this and all the AI stuff with real brain observations and computational neuroscience. This author is brilliant!
the best explanation I've ever come across, either in statistics books or on the Internet. You have extraordinary didactic skills. Thank you very much.
The skeptic in me wants to say I understood most clearly from this lecture because I had a solid subconscious understanding from watching probably a hundred of these over the years
The other part of me tells the skeptic to shut up and appreciate the magnificence of this presentation. This is a great example of data storytelling from my view.
Amazing. Perhaps the best PCA explanation video in the internet.
Data Science program + other PCA videos != Understanding of PCA. This explanation was amazingly clear and concise without feeling like I'm being coddled. Just subscribed!
Dear Sir, you are the best at explaining PCA!. I thank you very much.
I am yet to see a clearer explanation of PCA in my life! Thank you so much!
I hated eigenvectors and eigenvalues for 20 years because I thought they were useless.
Now, I have an appreciation for these concepts because of your video.
Thank you!
This is the best explanation of PCA in the whole UA-cam!Thanks so much!!
Masha Allah, sir. What a presentation! I have not encountered someone who could explain this topic in such a way that even a novice like me can grasp it.
I was loitering here and there from one UA-cam channel to another to get a better explanation of algorithms, and finally I landed up here...sir, your explanations are so easy to understand,thank you for your efforts ❤️
I went over at-least 30+ videos + materials , articles from 2 weeks , At last you made it so clear . You are wonderful teacher 🙏
the best PCA video on the internet. i always struggled with eigen value and eigen vectors thanks man
Finally someone telling me why. why do we use eigenvectors of covariance matrix to calculate the Principle components. Thank you so much!
This is by far the best explanation I have ever seen on this topic, trust me I tried so many videos including university courses ❤
Most underrated tutor !!! He just explained PCA like ABC.
One of the cleanest and very simplest explanation of pca. Great job sir
I came here after watching Mit lecture. This is how it should be taught.
FEELING SATISFIED
TEACHERS LIKE YOU CAN TEACH ANYTHING TO ANYONE NO MATTER FROM WHICH BACKGROUD YOU ARE.
One of the best explanations ever, PCA or no PCA!
Just amazing. I have an exam and I have gone through so much material already but nothing could have summed PCA better than you did. Thank you!
One of the best explanation to PCA on youtube , amazing video
Now, I can keep the PCA idea in my head without forgetting it. Thanks, the best PCA tutorial on YT, excellent.
Thank you very much! I hardly comment on UA-cam but when you see real skills, you have to acknowledge. I have watched several other videos on PCA but didn't really get what was going on. This video is so simple and clear for anyone with elementary knowledge in geometry and statistics to understand. Thank you very much !
The covariance explanation just struck me unguarded, it blew my mind. Thank you