Scribd
Upload
283 views

Summary of Formulas About Simple Linear Regression

This document provides formulas for simple linear regression. It describes how to calculate the coefficient estimators (b1 and b0) using the method of least squares. It also provides formulas for residuals, explained variance (sum of squares due to regression, SSR), residual variance (sum of squared errors, SSE), total variance (SST), and the coefficient of determination (r2). Confidence intervals and hypothesis tests for the slope coefficient b1 are presented using pivotal quantities based on the t distribution.

Uploaded by

William Noguera
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
283 views

Summary of Formulas About Simple Linear Regression

This document provides formulas for simple linear regression. It describes how to calculate the coefficient estimators (b1 and b0) using the method of least squares. It also provides formulas for residuals, explained variance (sum of squares due to regression, SSR), residual variance (sum of squared errors, SSE), total variance (SST), and the coefficient of determination (r2). Confidence intervals and hypothesis tests for the slope coefficient b1 are presented using pivotal quantities based on the t distribution.

Uploaded by

William Noguera
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

pag.

Simple Linear Regression Formulas

Coefficient's estimators with method of least squares:


SxY (xi − x̄)(Yi − Ȳ ) xi Yi − nx̄Ȳ
b̂1 = = = b̂0 = Y − b̂1 x
Sxx (xi − x̄)2 x2i − nx̄2

Distribution of b̂1 and b̂0


1 2 x2i 2
b̂1 ∼ N (b1 , σ ) b̂0 ∼ N (b0 , σ )
Sxx nSxx
Residuals
Yi − ( b̂0 + b̂1xi) = Yi − Ŷi

Residual variance or sum of squared errors


SSE = (Yi − Ŷi)2

Explained variance or sum of squares due to regression


n 2
SSR = Ŷi − Ȳ
i=1

Y Variance
SY Y = SST = (Yi − Ȳ )2 = Yi2 − nȲ 2
Formulas for SSE e SSR
2
SY Y Sxx − SxY
SSE = = SY Y − b̂1 SxY SSR = b̂1 SxY
Sxx
A formula for SY Y = SST
SY Y = SST = SSR + SSE
Coefficient of determination r 2 and correlation r
2
SSR SSE SxY SxY SxY
r2 = =1− = b̂1 = r= √
SY Y SY Y SY Y Sxx SY Y Sxx SY Y
SSE
Distribution of
σ
(Yi − Ŷi )2 SSE SSE
2
= 2 = MSE ∼ χ2n−2 ; E = σ2
σ σ n−2

Standardized of b̂1
b̂1 − b1
√ ∼ N(0, 1)
σ/ Sxx
Using the symbology of statistical software R, then

SSE SSE
σ 2 Σx2i (n−2) Σx2i σ2 (n−2)
se(b0 ) = = Sxx se(b1 ) = = Sxx
Sxx n n Sxx

σ 2 Σx2i σ2
se(b 0) is the estimation var(b0 ) = and se(b 1) is the estimation of var(b1 ) =
Sxx n Sxx
pag. 2

Pivotal quantity for b̂0

b̂0 − b0 (n − 2)Sxx n
= (b̂0 − b0 ) ∼ tn−2
se(b0 ) SSE Σx2i

Pivotal quantity for b̂1

b̂1 − b1 (n − 2)Sxx Sxx


= (b̂1 − b1 ) = (b̂1 − b1 ) ∼ tn−2
se(b1 ) SSE M SE

Hypothesis test on b1 of significance 1 − γ


Statistical test
(n − 2)Sxx (b̂1 − β)
(b̂1 − β) ∼ tn−2 OR
SSE se(b1 )
H0 : b1 = β; e Ha : b1 = β; critical region

1+γ SSE 1+γ SSE


b̂1 < β − tn−2 ( ) β + tn−2 ( ) < b̂1
2 (n − 2)Sxx 2 (n − 2)Sxx
or

b̂ β − tn−2(1 + γ ) se(b )
1< β + tn−2 (
1+γ
) se(b1 ) < b̂1
1
2 2
H0 : b1 ≥ β; e Ha : b1 < β; critical region

SSE
b̂1 < β − · tn−2 (γ) OR b̂1 < β − se(b1 ) · tn−2 (γ)
(n − 2)Sxx

H0 : b1 ≤ β; e Ha : b1 > β; critical region

SSE
b̂1 > β + · tn−2 (γ) OR b̂1 > β + se(b1 ) · tn−2 (γ)
(n − 2)Sxx

Confidence Intervals of level γ for b1.


Two-sided:
1+γ SSE 1+γ SSE
b̂1 − tn−2 ( ) < β < b̂1 + tn−2 ( ) .
2 (n − 2)Sxx 2 (n − 2)Sxx
or
1+γ 1+γ
b̂1 − tn−2 ( )se(b1 ) < β < b̂1 + tn−2 ( )se(b1 ).
2 2

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy