Regression Mathematics
Вставка
- Опубліковано 17 лют 2019
- Everyone needs to understand regression! Its a useful data science technique that allows us to understand the relationship between different variables. In this video, we'll play the role of a newly hired data analyst at a genetics company trying to find the relationship between advertising mediums (TV, newspaper, radio) and ticket sales to our newly opened theme park. Along the way, we'll learn about 5 types of regression models (linear, non-linear, multiple, lasso, and ridge). Expect math, code, and layers of explanation. Enjoy!
Code for this video:
github.com/llSourcell/ISL-Rid...
Please Subscribe! And Like. And comment. Thats what keeps me going.
Want more education? Connect with me here:
Twitter: / sirajraval
instagram: / sirajraval
Facebook: / sirajology
More learning resources:
• How to Do Linear Regre...
www.analyticsvidhya.com/blog/...
statisticsbyjim.com/regression...
hbr.org/2015/11/a-refresher-o...
blog.minitab.com/blog/adventur...
Join us at the School of AI:
theschool.ai/
Join us in the Wizards Slack channel:
wizards.herokuapp.com/
Please support me on Patreon:
www.patreon.com/user?u=3191693
Signup for my newsletter for exciting updates in the field of AI:
goo.gl/FZzJ5w
Hit the Join button above to sign up to become a member of my channel for access to exclusive content! Join my AI community: chatgptschool.io/ Sign up for my AI Sports betting Bot, WagerGPT! (500 spots available):
www.wagergpt.co
"Microsoft run by UA-cam sensation Bill Gates" lmfao dead
Love the interpretation of the regularization term as introducing bias to reduce variance! All the other people so far explained it as a penalty for having high weights, which is intuitive and nice to understand but do not include this second aspect of a bias term. Thank you for this insight, Siraj!
This is my entire base for starting Data Science :D
7:50 How exactly is the graph in the middle an example of a linear regression model when the actual graph is not a linear function? curvature as mentioned later is the exact opposite of something that has a linear progression. ..or am I missing something important here?
I also have the same question.
Hi siraj, thanks for uploading this video, I have a question, in the end of the video, you said, if the y has high collinearity with independent, its better to use lasso or ridge regression, but how can I ensure I have high collinearity? Should I ensure hight collinearity through r2 ( r square) please advise me
Much respect to Siraj. Always well knowledge and versed on the subject matter
In EDA, if we find the collinearity between two features, whether dropping one of those features or combing them to a single feature and using normal regression techniques helps? Or, is it necessary that we should go for LASSO or RIDGE regression?
I was there for your session in Mumbai. You were so amazing
Regression is a powerful tool for forecasting. Economists using it successfully predicted ten out of the last two recessions ...
Does that mean they are not factoring in all the variables (of which there might be trillions as anything can affect the economy worldwide)?
@@catalepsy8916 The number of variables that are needed to be processed for a perfect prediction are beyond our computation power as of now ..
Indeed
@@lugrisa R, STATA, MATLAB, GRETL, eViews, there are a lot of software solutions for econometrics nowadays, that it's actually hard to choose "correct one". IMHO Learning Python seems most reasonable for me nowadays, thanks to sheer amount of libraries
I need to know to predict dataset like csv to make machine learning.... With sklearn I didn't succeed to make It when I import and load my dataset with dataframe like df=pd.read_csv("")..
I need little help... thank you so much for your request..
I need to use sklearn to make machine learning.
I'm working on a dataset for a predictive analytics project around 5 GB. Preprocessing them takes up all the time. Is there a way to parallelise and speed up the process?
Always eager to see your videos 💖💖
Do you have a vid buidling on one of your two regression vids for multi-variable regression?
Best video on regression!
Is there a closed form solution of ridge regression with the non-negativity constraint of output variable?
Great video, Siraj. Would you be sometime doing a video tutorial on Plotly?
Hey siraj, Can you create a video which tells how to actually create a dataset from received signals? There is not much documentation given for dataset using signal processing. Thanks.
How can we work if we have to apply a linear regression in a complex-valued dataset?
Random Forrest is it powerful model to regression problem? @Siraj_Rawal
Thanks Siraj!
I prefer to use Random Forest, Support Vector Regression and Decision Tree Regression to solving my regression problem sometimes Polynomial Regression is also giving a good prediction on a test set.
Great Intro :D
Siraj hits gym I guess so🤓🤓
At 3:14 you say that the dependent variable can also be called the predictor variable. I believe it's called the response variable instead. Predictor/explanatory/independent variables are the same.
great intro Siraj! Agree with Bookerer... predictors are the independent variables X
Yes, Predictors are independent variables(x1,x2,....xn). But, dependent variable is Predicted (Y)
I've seen some machine learning courses that used the term "predictor" for the dependent variable.
Notification : Hey, Siraj just uploa....
Me: Say no more...
From India?
@@ParmodKumar-fw7jy yep
GJ man! Can I get a heat?
I just made my expression of understanding
Fantastico
Nice explain
Hey Siraj! First to comment!
nice
I love this guy
I love lizard girl
Hey man, just wanted to let you know that I really appreciate your videos.
At 3:11 he means Y_i = \beta_0 + \beta_1 X_i instead of Y_i = \beta_0 + \beta_1 X_1 , where i index your couple (X,Y) of your dataset.
I was first to view the video :)
How lasso make irrelevant features to zero? I mean what is the mathematics behind it?
regression my fav
Can i Know Math behind the linear regression
Why lasso penalise the high coefficient to zero while ridge only makes it a reduced value?
I don’t know if you saw my other comment, but you are talking much slower now and your videos are way better!
I love how we are relearning high school math :)
Who wants to be a MLionaire
seasoned redditer spotted
@@VahidSaffarian what does that even mean? 🤔
Great video, please update the school of AI link!
The regression line at 5:24 does not follow the data at all! Although they would be so nice and linear...
When's your Meme Review with Elon?
Haha, ask him on Twitter! I’d love to this week
1st comment and 1st like
Just explained how to do my senior assignment better than my teacher.
You are great sir, inspiring and improving skillset the fun way.
7:00 "Does this equation explain the meaning of life?" 😁 Another great video. It's good to have these high level overviews of a topic (eg. linear regression) because they give a good framework to build on once you start studying them in depth.
The question is did you use regression on pop culture references to write the script ?
There are another types of regression model that you didn't consider inside your list, for example: non linear, semi parametric, generalized linear models, additives models an so on
For anyone interested, I wrote a few of detailed python notebooks on linear regression and also ridge + lasso a while ago:
Linear Regression Notebook + PDF note (applied to Fifa 2018 data):
github.com/hammadshaikhha/Math-of-Machine-Learning-Course-by-Siraj/tree/master/Linear%20Regression
Ridge Regression (applied to IMDB data):
github.com/hammadshaikhha/Math-of-Machine-Learning-Course-by-Siraj/blob/master/Regularization%20in%20Linear%20Regression/Regularization%20in%20Linear%20Regression.ipynb
Lasso and Ridge Regression for model selection (applied to NY school data):
github.com/hammadshaikhha/Math-of-Machine-Learning-Course-by-Siraj/blob/master/Linear%20Model%20Selection/Linear%20Model%20Selection.ipynb
I wonder, how can people even give dislike on this video! Dumb people :/ They don't know how to appreciate good work.
Simple Explanation to Regression
'Siraj's Quality Content'
--- Linear Regression
BST 210 squad up!!!
1st like
the best weapon in financial trading to destroy the market makers ;)
Is Siraj's voice unusually deep here?
May be it is AI powered voice ..😀😁
Hi Siraj,
Thank you for this video.
i want your reply on this. I am studying ML for a long time but unable to crack the interview. i try to follow your 3 months ML curriculum. but unable to understand what to do, where to work. where to do practices i dont know.
please help me on this. Thanks!
when u r studying for long time, then why r you so confused.
Make 1-2 portfolio projects.... like for self-driving car, swarm intelligence etc.
@@DeependraTube I dont know how to start, plz share some reference links.
@@AshishTyagi2911 Do u know Coding already ?
If so , you can go to fast.ai and search youtube for Andrew Ng course , and also Coursera have good courses.
When a tokai learns to print hello world!!
Yo
Early
First
Second
It’s funny how the video is entirely about linear regression and I didn’t hear the word “correlation coefficient” one time.
those hands 😂
404 ERROR: RAP SONG NOT FOUND
Really... Fortnite..
Any video with Math on it's title gets auto-downvoted instantly by a script I'm running on background. Crap!
waste of time