Multiple linear regression - explained with two simple examples
HTML-код
- Опубликовано: 3 сен 2021
- See all my videos at www.tilestats.com/
In this video, we will see how multiple linear regression is computed. The focus is to understand how to interpret the coefficients in models with both categorical and numeric variables, and models with interaction terms.
Alas! I found a video I could relate with well when it comes to multiple linear regression. The examples you used gave more clarity. Thanks so much
i've been looking for this example! so clear and well explained. thank you!!!
Thank you!
Best channel❤ but underrated ..
Simple and intuitive explanation of complex concepts..
❤
Thank you! I finally understood how to interpret the coefficients in a multiple linear regression model.
Great!
thank u for a good explanation
Great explanation for interaction term. Thank you.
One question though. Here you explained the meaning of interaction between a continuous variable and a categoric variable. How can we interpret the interaction when both the terms in the interaction are Continuous variables and when both the terms are Categoric variables?
Well explained. Thanks a lot!
Thank you!
Very good.
Thank you
Can I ask... Where did you get that 30.57 and 3.55?
Have a look at this video
ruclips.net/video/taPvVyJVc_A/видео.html
if I had a category such as car dealer which has more than just two options (so I can't just put 0 and 1) how would I go about incorporating that?
say we had dealer a, dealer b and dealer c, where the difference is noticeable between them
Thank you for this. What if we code Gender as 'M' and 'F' and not 0 and 1. Then at 10.50, it will be 12.9xM and not 12.9x1. Then how can we include 12.9xM in the intercept? What I mean is that 0 and 1 in this case are factors, can we multiply 12.9 with a factor (treating factor as a numeric)?
Also we can choose any other number instead of 0 and 1 e.g. 3 and 9. Then in this case the intercepts will be different. So, this seems arbitrary as the Gender intercepts depend on the way we choose the numbers?
You should always use 0 and 1 to recode a categorical variable because they represent absent or present. However, if you use a software, it will do this automatically.
Which software is used to get the equation for model
Price = constant + Age.Coefficient + Mileage.Coefficient ?
You have to create the equation on your own and then use the software to estimate the parameter values for the equation. I use R but you can use any other statistical software to get the same parameter values.
This is a great video.
Does this mean we are actually analyzing men and women differently as we get two different regression lines: one for men and one for women. How will this compare if we run the model stratified by Gender?
Yes, men and women are predicted differently by the model because they have separate regression lines.
Why do you change the example midway? You didn't explain how you calculated the equations.
do you have the data to to solve the coefficients in your example ?
At minute 2:18, 7:03 and 12:38 you have all the data to reproduce the results, including estimating the coefficients.
I still can't get it where the data is where you calculate the b0 and b1 and b2
I do not show how to calculate the coefficients by hand. I simply plug in the data in a statistical software to compute these. In this video:
ruclips.net/video/taPvVyJVc_A/видео.html
I show how to calculate parameters by hand, but only for simple linear regression. For multiple linear regression it is a bit more calculations and I currently do not have a video on that.
but how to cumpute coefficients in multiple regression?
Use a software, or if you must do this by hand I would recommend this page:
www.statology.org/multiple-linear-regression-by-hand/
I also have a video on OLS for simple linear regression:
ruclips.net/video/taPvVyJVc_A/видео.html
How is the code “0” or “1” determined? What if you had a third category? Would it be “2”? Only part that I didn’t follow fully.
No, then you add an extra term in the equation. If term 1 = 0 and term 2 = 0 it will represent the baseline group(group 1) If term 1 = 1 and term 2 = 0 it will represent group 2, if term 1 = 0 and term 2 = 1, it will represent group 3.