Logistic Regression - VISUALIZED!

[@ulrikeschnaithmann3914]

This visualization is so strong, I feel like it's like one of those lecture in university that were so good, you'll never forget them!

[@CodeEmporium]

Thank you for the amazing compliments! I definitely tried a lot with this one haha

[@iyar220]

I'm currently self studying machine learning, and the very first thing I did after learning about decision boundaries was check if 3blue1brown had a video on it, because his animations are just incredible for understanding math and gaining intuition. The way this video is constructed, and again, the animations used in here are incredible. Thank you so much for making this!

[@CodeEmporium]

Glad you liked the visuals here. Honestly creating this video solidified my understanding of this as well :)

[@adityagitte]

This video is exactly what I was looking for, I was getting confused between decision boundry and the sigmoid function for a 2d input problem, thanks a lot for the wonderful animations!

[@matthiasmitchell4801]

This is insanely helpful. I feel like I can actually use logistic regression libraries and have a good idea of what's happening. Generalizing this to higher dimensions was eye-opening--I never quite knew what was going on there.

[@alexanderdellorti6039]

Thank you mate, I think visualization is so underestimated in universities, in math and science THIS is the key of teaching. Great job!

[@CodeEmporium]

Thanks for the kind words. I thought this video didn't receive the attention I thought it should have. But glad there are others out there like you who find value here. :)

[@mysillyusername]

A brilliant visualization of the logistic regression, thanks for making this!

[@CodeEmporium]

My pleasure :)

[@hypebeastuchiha9229]

This video deserves a million views

[@RiteshSinghArya]

Beautifully explained

[@bhuvandwarasila]

That was so fire! Nice bro!

[@jose4877]

This was very cool! Thank you!

[@arpitsahni422]

simply lovely mate! this helped me connect everything

[@olusanyatodd4083]

Oh wow! This is so great! I just learned the math in a class but this really explains the intuition! Thank you so much

[@CodeEmporium]

Of course :) Thanks for watching

[@aniketchhabra8912]

This is amazing!! Thanks a lot for sharing

[@BiranchiNarayanNayak]

Excellent explanation of Logistic Regression

[@apoorvshrivastava3544]

dude please be regular You will earn more subscribers as you desrve million of subscribers keep it up

[@CodeEmporium]

Thanks a ton! I had some life changing events take place in the last few months (graduated, travel, moved, new job). Now that things have settled a bit, I can be more regular :) Thanks for the support. Means a lot!

[@darasingh8937]

Great video! Thank you for your time and creativity!

[@Leibniz_28]

17:32 ¿"m" dimensions or "d" dimensions? Great video, your content is really high quality

[@CodeEmporium]

Yup. I didn't write m because I didn't want to confound this with the "m" in number of iterations.

[@clearwavepro100]

Thank you! Super helpful on a lot of levels :)

[@taiworidwan194]

Thanks for the video. It is really helpful Please, how can one optimize the coefficients of a Logistic Regression Model using a Genetic Algorithm?

[@amoughnieh]

Thank you so much for this video!

[@CodeEmporium]

You are very welcome. Thank you for watching

[@emilycoppens4603]

This video was great, thank you

[@Tntpker]

The visualization at the end is what wee need more off

[@CodeEmporium]

Glad you like the visuals!

[@romanmelnyk1777]

Why decision boundary doesn't depend on activation function? So if i want to have a curved dicision boundary, i don't have to change activation function, but rather change my features?

[@shubhamsingh6884]

Great video. It really helped me to get a better understanding of logistic regression. However, I have a couple of queries - What is the target function of logistic regression which s being learned (like in linear regression we have y = w.T*x)? To what curve do we fit the training data, the decision boundary of sigmoid function (like in linear regression we fit the straight line defined above)? Thank you !!!!

[@mohanakumaran5815]

I really love this channel 😘 Pls post many videos often, not once in blue moon 😁

[@CodeEmporium]

Super glad you do! Had some life changing events come in the last few months (graduation, move, new job). But now that things have settled, I'll be more frequent. :)

[@soryegetun529]

awesome explanation thanks so much

[@CodeEmporium]

You are so welcome

[@bytesizebiotech]

Love your stuff. I'm not a math major, and I've learned that you don't have to be to understand, but it's kind of offputting when people use a ton of symbols. It's not necessary to explain what is going on

[@urfiyogabama2589]

Great!! Thank you

[@skewbinge6157]

thank you so much

[@PremKumarAmanchi]

Can u make a video which explains the plot for 3 labels using 2 features, I am just curious how the sigmoid function looks for it. Amazing work!!!

[@CodeEmporium]

Thanks so much for watching! I’ll keep your suggestion in mind and see if there is leeway to do this at some point (tho I don’t think it will be anytime soon admittedly).

[@X_platform]

These visualizations are hot! Love both of your channels :)

[@kpratik41]

which is the other channel?

[@X_platform]

@@kpratik41 3blue1brown

[@muhammadusman5521]

Please explain how the value for bias and weigh are calculated That sigma there If n is also 1 Then y1 and x1 What are there values Please I want to calculate and loop them like you did

[@patite3103]

You've done an amazing video! At 14.40 you show a plot in 3 dimensions. On the z-axis (vertical one) the values should spread on the interval [0,1] which is not clear here. It's not clear why the boundary is not a plane since we have a scatter plot in 3 dimensions. I'm quite confused with the dimensions of the plot.

[@josephfdunphymba3241]

not seeing a link to visualization program

[@tawhidshahrior8804]

DUDE YOU ARE A LIFE SAVER. Subbed and reccommended to my fellow msc colleagues. Keep up the great work brother.

[@CodeEmporium]

Thanks a ton for the share :) And so happy this helps

[@kdang7233]

why do we need to use e^-x instead of any f(x) >=0 with every x to express the possibility, I forgot all the math learned from highschool so At This Point I'm Too Afraid to Ask

[@javxa]

Oh boy, this visualization is incredible. I always knew there was a sigmoid function in 2D. Guess what... It was hidden in 3D LOL

[@CodeEmporium]

The mystery has been solved by Detective Emporium.

[@ermiassolomon7605]

what is the book in the beginning of the video?

[@CodeEmporium]

That wasn’t a book; just me listing out some topics :)

[@vijayendrasdm]

loved the explanation. But, why do we use sigmoid function ?

[@balajikannan7393]

A sigmoid function transforms any real number to be between 0 and 1. In other words probability would lie between 0 and 1.

[@vijayendrasdm]

@@balajikannan7393 There are thousands of function that can map real number to a number between 0 and 1. For example Step function being one of them. Then , why pick sigmoid out of hat ?

[@jeverly]

@@vijayendrasdm sigmoid function is differentiable which allows us to train our weights using gradient descent, step function is not differentiable

[@amarnathjagatap2339]

Sir it's amzing make more on ml

[@Simon-mv6zn]

9:03

[@viddeshk8020]

Bro, please use dark theme

[@CodeEmporium]

I do now. Recent videos