Binary Logistic Regression Tutorial

Binary logistic regression is a machine learning algorithm most useful when we want to model the event probability for a categorical response variable with two outcomes (yes/no, true/false, etc.). In this video we'll build a sentiment classifier app that uses binary logistic regression to classify tweets as either happy, sad, or neutral. I'll use animations, code, rap, skits, and equations to explain how it all works. Enjoy!

Code for this video:

Please Subscribe! And like. And comment. That's what keeps me going.

Want more education? Connect with me here:

This video is apart of my Machine Learning Journey course:

More Learning Resources:

Join us in the Wizards Slack channel:

Sign up for the next course at The School of AI:

And please support me on Patreon:

Original repo:
Signup for my newsletter for exciting updates in the field of AI:

[Artificial_Intelligence_AI]

Great video as always but just a clarification: You don’t need logistic regression to determine if a tweet from Kanye makes sense...

That’s a constant with always the same value of 0.

[rishisharma]

I think some of the slides are from the course of Machine Learning by Andrew Ng

[vgtgoat]

Omg I just spent the last 2 weeks at my internship trudging through logistic regression articles online to build a financial risk scorecard model. Where were youuuu.
At least I got a great review out of this thanks for the video!

[postnubilaphoebus]

Hey Siraj,
I just wanted to tell you that you're awesome! I'm a psychology student but I'm currently writing my thesis on CNNs. Your Videos helped me a lot to understand the subject matter intuitively, despite the fact that I don't have a strong math background.

[VuqarIE]

Dude, I love your videos, Grammy or Oscar awards should be given for your videos.

[megathai]

Time to go to sleep, it’s 3:26 AM

*Siraj uploads a new video*

Well damn. I will going to bed at 3:37

[wolfisraging]

U r the only best source for dl and ml on the planet

[KshitijBajracharya]

The rap wasn't so good Siraj. "Low error, pop champagne" is still the best :D
Jokes aside, always learnt a lot from you. Can't thank you enough.

[scott]

Nice vid. We used multinomial logistic regression models at work for our IFRS9 models to model probabilities of customers moving through different credit states.

[RandomGuy-hijm]

*U made that very basic topic a very interesting one*

[dineshpabbi]

I have started learning machine learning so since i have practiced many things, i loved this video of yours because i understood everything so well 😂

[jomsbryan]

Thank you Siraj! I was looking for a best model to use for Sentiment Analysis and found your video very helpful. I just subscribed and hit that bell notification.

[ashwinpande]

Nice one! Btw you never notice my comments although I've been watching your vids for a long time. You're also one of my favourite youtubers. Quite a busy man haha.

[nxbil]

Hello Siraj, i love all of your video but please improve by put a subtitles because sometimes it is more better to understand.I hope u notice my comment😊🙏

[soumya]

Awesome dude.. And u look so funny when i listen u raps in 3x... 😂 😂 😂 Keep it up...

[rahulbera]

great video .... will be waiting for the video on multi class classification....as i get very confused while choosing a algorithm for multi class
(don't choose SVM they are fu*k*ng slow)

thanks a lot siraj...keep bringing videos like

[asadullahfarooqi]

like i am confused about the types of Decision trees (id3, info gain ... etc) which should to be used and why it's math make sense .... it would be very helpful thanks

[xPROxSNIPExMWxPOWER]

U should make a video on Spiking Neural Networks.

[marcwilliams]

Could you turn on captions please? Thank you!

[thiagoleite]

Hi Siraj, if you can, please enable Youtube's automatic CC. Thanks a lot.