free web page hit counter
🛡️
Copyright Notice: This video is officially sourced and embedded from YouTube. For all copyright inquiries, reports, or removals, please contact YouTube's legal team here.
Luis Serrano Academy

Luis Serrano Academy

194,000 subscribers

⏱ 👁 463,949 views

Principal Component Analysis (PCA)

Video Overview & Insights

Announcement: New Book by Luis Serrano! Grokking Machine Learning. bit.ly/grokkingML

“

The best introduction to PCA on YouTube I've seen so far.

— @ensys1000

40% discount code: serranoyt

A conceptual description of principal component analysis, including:

“

I watched like 10 videos and this is the only one that explain the PCA in a simple and detailed way . Thank you .

— @mustafaabdalla4309

- variance and covariance

- eigenvectors and eigenvalues

“

Tu video me lo recomendó ChatGPT, que bien explicado, increible!

— @sergiofernandeztesta6433

- applications

As usual, very little formulas, lots and lots of pictures!

“

very good explanation ! finally I got it ! 🙂

— @kawabiker2655

0:00 Introduction

0:46 Taking a picture

“

Great work, You have spend lot of time to make the PCA concept as simple as possible. Keep Going. I make use of this content for my future work by acknowledging you in first slide. Many many thanks Prof.

— @anbarasup1293

1:13 Dimensionality Reduction

2:02 Housing Data

“

Incredible video sir! I hope you do one about factor analysis and MANOVA. What you did in this video is extraordinary

— @sterian1673

5:09 Mean

7:46 Variance?

“

superb Video..the best ever explanation about PCA

— @ambudevaraj4706

12:47 Covariance matrix

13:58 Linear Transformations

“

Wonderful !

— @DertiDerty

18:12 Eigenstuff

19:16 Eigenvalues

“

This is the most beautiful explanation of PCA I’ve ever seen

— @irksomevalkyrie

19:53 Eigenvectors

20:51 Principal Component Analysis (PCA)

“

Imagine paying thosands of dollars for university where they provide over-complexed matriials and explanations, while this video explains entire chapter in less than half an hour. Thank you very much!

— @arrakis3

26:05 Thank you!

“

Combining theory with example, step by step. Love the clarity and simplicity. 👏

— @mitakar6360

More User Perspectives

@

Perfectly complete explanation!

@EvansCodingCorner
@

Thank you for your video, you explain the concept of PCA very clearly❤

@ahmadriza8909
@

So impressed! This channel is a hidden gem! Thank you so much.
PS: Make sure to watch the video till the end, where everything is wrapped up.

@hosam_rahma
@

2026, who is here?!!!! Lovely explanation !!!!

@STANS2447
@

More than amazing ❤👏!

@Abdirahman_Halane
@

Best explanation so far to PCA on Youtube!

@drachenschlachter6946
@

excellent! thank you very much dear friend

@stefanomariani2454
@

at 4:48, why do you call the dimensionality reduction to size instead of number of rooms, the x axis was representing rooms in 2d scatter plot, wasn't it?

@rakeshgunijan5401
@

this channel's videos are truly the best for its kind. highly recommended

@ankurbaral1543
@

AMAZING

@sathwik7362
@

Thank you.

@AdamMason-t7c
@

perfect explanation! thanks a lot

@byzbutterfly
@

Excellent presentation

@JindaJatt-d4q
@

This step by step explanation starting from the mean concept and variance, then moving through the covariance is a great idea that answers my question of why we use the covariance here. Big thanks

@MedhatR-do9le
@

Thank you bro, I have been researching about this but I havent understand any at all until I watched this video

@hyperkarlotheweeb2969
@

Brilliant, no words.

@code-at-amrita
@

this is probably the best video ive watched in my life over 1000s of math videos. Instant sub

@neeldesai501
@

00:50 I really like this analogy :>
06:48 What if the numbers are complex? Then squaring them could still give a negative number, right?
08:00 Couldn't we calculate the centroid of this cloud of points, and then calculate the squared distances for each point in the cloud from this centroid, and divide by the total number of points? That would be a single number that characterizes the "overall spread", wouldn't it? If this isn't the variance, then how is this called?
19:30 Which might be nice enough for 2×2 matrices, but with 3×3 and bigger you may not get that lucky, because in most cases you'll end up with some crazy cubic/biquadratic/quintic/something worse that cannot be easily solved algebraically. And if your matrix is 100000×100000, then forget about calculating the determinant (which has exponential complexity) :q Definitely not feasible for practical applications. I suppose something like singular value decomposition might be more feasible in that case.

@bonbonpony
@

❤❤❤very understanding

@nafeesathraffa8532
@

Good Analalysis,Thank You

@pprasad6937
@

3 years after learning linear algebra i finally understand the meaning of eigenvalues and eigenvectors

@voltaire8828
@

❤❤❤❤🎉🎉🎉🎉

@奶爸博士
@

I have never seen such a simple and clear and life time remembering explaination of PCA

@SakshatBagadi
@

Absolute best explanation on YouTube! 🎉 subscribed

@andreaurdaneta3835
@

Great explanation!

@snargundkar
@

only thing i was not able to get in PCA was what happens to the other eigenvector with smaller magnitude. Now i know , it gets the boot. will sleep peacefully

@nssandhu4401
@

One of the best video to build the intuition of PCA !! Thanks a lot for putting this together !!

@RahulBhagatclicker
@

Very nice explanation with perfect examples.

@kibriahossen1379
@

you beat StatQuest on this one - must say!

@achm3dz
@

was perplexed what was PCA ? now i have got my answers and thank you for such practical elucidation

@RishavPatel-ok2mh
@

this clear my confusion abt linear algebra so much

@yangsong6111
@

The best explanation of PCA on the YouTube till now compared to the best channels playing there...

@praneethaluru2601
@

honestly the BEST ONE

@maniksh3204
@

Excellently explained. Thank you very much.

@spontinimalky
@

3blew 1blown mentioned 🗣️☝️🔥

@masterplayer5982
@

I couldn't understand the part where you got 5 eigen vectors but you took just the first two after sorting in descending order of eigen values, why didn't all the five eigen vectors can be used?

@rochittupe
@

Perfect

@Rauf_Akbaba
@

very well explained. thank you!!!

@ianchong7422
@

Thank you for this awesome video. I like it for a fact that particularly the covariance was never been explained this clearly by anyone.

@mend-amarbadral5583