Regression 1 : What is regression? | SSE, SSR, SST | R-squared | Errors (ε vs. e)

The first video in a series of 5 explaining the fundamentals of regression.

Ever wondered WHY you have to SQUARE the error terms??

Here we deal with the very basics: what is regression? How do we establish a relationship between two variables? Why must we SQUARE the error terms? What exactly is SSE, SSR and SST? What is the difference between a POPULATION regression function and a SAMPLE regression line? Why are there so many different types of error terms??

Enjoy.

Category:

Howto & Style

Tags:

• What is regression
• R squared
• SSE
• SSR
• SST
• Statistics
• Error terms.

[ejgi]

and thanks for sharing!

[sachin01663]

@juzzzeltzer ~ Thanks a lot.

[juzzzeltzer]

@sachin01663 Hey Sachin. You gotta be careful with the sums of squares. I've used

SSR = Sum of Squares to due Regression.

SSE = Sum of Squared Errors.

Some textbooks use SSR = Sum of Squared RESIDUALS. Which would be the same as my SSE. It sounds like that wiki definition is doing that! So yep, it's a little confusing when different sources use different abbreviations, but you'll find that happens ALOT in stats. So just be careful to define your acronyms when you use them.

Hope that helps

[sachin01663]

First of all, very well explained. Thanks a lot. Small doubt--

SSR and SSE concept is little confusing.

According to wikipedia - Explained sum of squares and SSR is not same. SSR is a measure of the discrepancy between the data and an estimation model.

However, according to your video, its the Squared distance between 'Mean value (not actual) and Predicted Value'. SSR and Explained Sum of Squares is same.

Any explanation will be very useful.

Sachin -- sachin01663@gmail.com

[julsiebear]

Love it. Many thanks!

[nyczcrzyazn1]

You explained the concepts better than my professor! Thank you!!

[StanleyMT]

Awesoooomeee. I have an exam in 13 days an this helped a LOT