Statistics 101: Linear Regression, Residual Analysis
HTML-код
- Опубликовано: 5 окт 2024
- In this Statistics 101 video, we learn about the basics of residual analysis. To support the channel and signup for your FREE trial to The Great Courses Plus visit here: ow.ly/xVD030fiZ8S
So this video is the next in my series on Simple Linear Regression, however, I will say that this video does have implications beyond just simple regression. So this video is about residual analysis. The residual analysis will do two primary things for us. Number one, it will tell us how well the model we have produced fits the data we are looking at, or in other words, how is our error? Is our error large or is our error small? Number two, maybe most importantly, it will tell us whether or not the model we are using is actually appropriate for the data we are looking at. So as you know, probably by now, there are many, many ways to model a data set and there are certain models that are more appropriate than others and our residuals can help us decide that. So residual analysis goes well beyond just statistics, it goes into higher-level statistics, it goes into data science, and of course, machine learning when we're talking about which model we should choose for our application.
So if you are new to Regression or are still trying to figure out exactly what it even IS...this video is for you. Sit back, relax, and let's go ahead and get to work.
My playlist table of contents, Video Companion Guide PDF documents, and file downloads can be found on my website: www.bcfoltz.com
#statistics #regression #machinelearning #datascience
This is easily the best video series I have seen on Linear Regression and has provided me with a fantastic introduction to the subject area. It speaks volumes of your teaching ability and the quality of this series, that this subject area which may be perceived as overly difficult from the outside looking in, has been taught in a way that makes it accessible to basically any and everyone. I tip my hat off to you Brandon and thank you very much for your time and effort. Not only do I feel confident for my test next week, but I know that moving forward while building up my knowledge base, I can use your videos as a reference point. Now onto Multiple Regression :)
Agreed and well said!
Couldn't agree more!
same, dude is very good at teaching stuff
I can't say how much I appreciate this video. Simple, elegant, and well-explained. I had a hard time understanding linear regression in my quantitive statistics class, but I figured it out after watching this series. If only all the teachers out there are capable of making it easy and simple like Brandon.
I am so grateful for this series! I have taken many courses on stats, but this is the first time I feel like I really understand it. I wish I had teachers like you at my university. I hope you will do more videos in the future maybe about more advanced topics as well, like mixed models, cox proportional hazard models, mixture models etc.
Finally I've finished those tutorials for simple linear regression. Thank you so much for sharing the knowledge. Here is one little suggestion: It will be much better if you could spend less time on caculation of those numbers and Do pay more attention on explanation of the meaning of every symbols and formulas.
almost watched all of your non/linear regression videos, really helped me. thanks mate!
Thanks for another outstanding video. Brandon, you do a wonderful job of making concepts accessible. Thank you!
Hey Brandon, Hope u are doing great!! As we all are with your teaching. Can you please make some videos on the requirements or assumptions for the Procedures such as Linear/Logistic Regression, Hypothesis Testing with detailed study. Please make some videos on other popular tests also such as Fisher's Exact Test with Chi Square test comparison and so on... And for your Videos, Please keep them coming.... We all are enjoying the learning... Lots of love..
awesome .....could not get better this this
Hi Brandon, another terrific video! In your profile, you mention that you recommend looking at some books. Can you list them?
Hey Brandon, thanks for the great video. We enjoy a lot. Clear my basics
Great video, clear explanation. Thanks!
thank you man...love your style of teaching!
Another great video! Absolutely hooked to your channel Brandon
Hi,
from what I've read in other resources the SS_residuals is what you named SSE here. This could lead to some confusion.
Besides, great video!
Brandon, could you do a series on statistical softwares? I'm about to recommit to learning another but I have no idea what to use. I'd prefer open source. Should I commit to learning R or Python? Are others like SPSS easier with graphical user interfaces and possibly less coding? What tool is best for an analyst that only needs to do basic to moderate statistical analyses beyond Excel. It would be super helpful!
I don't quite understand the meaning of residual vs. y_hat. How it is different from R vs X and why we need that?
Hi - I agree fully with all the comments below. This whole playlist just cleared almost all confusion about linear regression that I have had throughout the years. Thanks for that. I'll make sure to like and donate. However, one little clarification would be nice.
So for example, Let's say I want to model an output of some product and it turns out that this can be fitted to a linear regression model. I have some quality criteria for this product that need to be fulfilled. As I wan't to be sure that the model explains the output or quality parameter in all future batches (based on the same production method and raw materials etc) I choose to set limits based on prediction intervals, as they explain how the individual values will come out with a probability of e.g. 95%. Would I then violate any assumptions? Can I predict the the mean in all future batch-samples using the confidence interval or is future prediction limited to the prediction interval? would I need to take anything into consideration for this. I hope my question make sense.
In the linear model, (y = beta-sub-zero + beta-sub-one etcetera) shouldn't there be an x included after the beta-sub-one, being the independent variable?
yes, he probably missed that.
Yes, I was about to comment to point that out. But you've already done it
At 5:55, the equation should read: "y = b0 + b1 * x + epsilon", but "* x" is missing.
THANK YOU!
In the 3 squared graph you've used y hat instead of using y bar. I think that's by mistake. Right??
I have also same comments.
Brandon you deserve to be a professor
The slide of the regression formula has an x missing (around 5:58)
Great video 👍 do you have one on regression using forecasting as the example?
Why is y-bar marked y-hat on the chart?
its a mistake
Hi thanks for this video! How would we report residuals in writing? Would something like, "here we report a residual of -0.56 between actual and predicted values" ?
Hello guys, what is the difference between Residual Plot against Y- hat vs Residual Plot against X ?
Thanks!
I had performed a linear regression and the residual value is 0.002, what does it indicate? whether the linear regression performed for the chosen independent and dependent variables were good or bad? and also residual is nothing but R-value?
@Brandon at 6:09 where you put state the regression model... shouldn't (beta sub 1) accompany x?
What is the sequence of these lectures
Where do we get conclusion that "everyd dollar increase in meal bill there is a predicted amount of $ 0.15 increase". Thanks
The slope is aprox. 0.15.
Why we saying errors will follow normal distribution
where did that sloppe 0.1462 appear?
the predicted tip being the mean.
SSR and sse are kinda swap.
T-H-A-N-K-S !!!
Brandon Foltz is not accomodating when it comes to questions.. I already posted around 12 questions in different videos and yet I haven't received a single response.
😢😢