0% found this document useful (0 votes)
281 views2 pages

QM II Formula Sheet

This document provides definitions and formulas for key concepts in regression analysis: 1. It defines the prediction equation, sample slope, sample y-intercept, coefficient of determination, standard error of estimate, and standard errors of β0 and β1 for simple linear regression. 2. It provides formulas for confidence intervals of β0 and β1, and for the mean value of Y given x, and discusses adjusted R2. 3. It introduces concepts like prediction intervals, variance inflation factors, tolerance factors, beta weights, and partial F tests which are used to evaluate multiple linear regression models.

Uploaded by

phanindra sairam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
281 views2 pages

QM II Formula Sheet

This document provides definitions and formulas for key concepts in regression analysis: 1. It defines the prediction equation, sample slope, sample y-intercept, coefficient of determination, standard error of estimate, and standard errors of β0 and β1 for simple linear regression. 2. It provides formulas for confidence intervals of β0 and β1, and for the mean value of Y given x, and discusses adjusted R2. 3. It introduces concepts like prediction intervals, variance inflation factors, tolerance factors, beta weights, and partial F tests which are used to evaluate multiple linear regression models.

Uploaded by

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

8.

Confidence Interval
of 0 and 1

REGRESSION
1. Prediction Equation

yi 0 1 xi

1 t( / 2,n 2 ) S e ( 1 )

0 t ( / 2 , n 2 ) S e ( 0 )

2. Sample Slope

SS
1 xy
SS xx

x x y y
x x
i

SSx= x2- ( x)2/n

9. Confidence interval
for mean value of Y
given x

12. Adjusted R2

SSE /(n k 1)
SST /(n 1)
n 1
RA2 1 (1 R 2 )
n (k 1)
RA2 1

A (1 - )100% confidence

interval for E(Y X ) :

SSxy= xy- x* y/n

Yi t / 2 S e

3. Sample Y Intercept

1 ( X i X )2

n
SS X

R 2A The adjusted coefficient


of determination

0 y 1 x
4. Coeff. Of Determination

R2

Backward Regression
Fout < 2.71
Pout > 0.10

SSR
SSE
1
SST
SST

Here Y is the E(Y X).

R 2 Unadjusted coefficient

10. Prediction interval


for a randomly
chosen value of Y
given x

of determination
n number of observations
k no. of explanatory variables
13. Variance Inflation
Factor
1
14.
VIF(X )

A (1 - )100% prediction
interval for Y is :

Yi t / s S e 1

5. Std. Error of Estimate

Se

n k 1

1 (Xi X )

n
SS x

where Xs are observed values


of independent variable.

Y is the estimate of Y, n is
the sample size and Se is the
standard error of Y
6. Standard Error of 0 and
1

S ( 0 )
S (1 )

Se

Se

(for simple regression only)

SS xx

Forward Regression
Fdefaultin > 3.84
Pin < 0.05

Estimate Parameter
Est. std. error of estimate

1 1

Se ( 1 )

SS XY
r R
SS XX SSYY
2

nSS xx

7. Test Statistic
t( n 2 )

11. Coeff. of Correlation

t( n 2 )

1 SSx
Se

1 R 2j

R 2j is the coefficient of
determination for the
regression of X j as
dependent variable

If VIF > 4, it is co
linearity
14. Tolerance Factor: 1Rj2
15. Beta Weights

Beta i

Sx
Sy

Sx Std dev of X
S y Std dev of Y
16. Partial F Test

Fr ,n ( k 1)

( SSER SSEF ) / r
MSE F

SSER - sum of squares of


error of reduced model
SSEF - sum of squares of
error of full model
r no. of variables
dropped from full model.
17. For Multiple
Regression:
A (1-) 100% PI (Prediction
Interval) for value of Y given
values of Xi:

19. Mahalanobis
Distance (for SLR)
Mi = [(Xi X)/ Sx]2
Outlier if value >10
Or can refer chi-sq table
20. Cooks Distance (for
simple linear
regression)

Cut-off: > 1 or 4/(n-k-1)

Semi-partial (or part) correlation


sr12,3 is the correlation between y1
and x2 when influence of x3 is
partialled out of x2 (not on y1).

sr12,3

Mahalanobis
distance

> Critical chi-square value


with df = number of
explanatory variables
(Outliers in independent
variable)

Cooks distance

> 1 implies potential outlier

23.

Leverage for a point


is given by (simple
linear regression):

Hat matrix relates the


estimated value of y (y-hat)
with the observed value of y
(thus the name hat matrix).
H=X(

The threshold value


for leverage is
2(k+1)/n or 3(k+1)/n.
Where, k is the
number of
explanatory variables
and n is the sample
size.
22.

Leverage
values

> 2(k+1)/n, then the point


is influential (k is the
number of independent
variables and n is the
sample size)

SDFBeta

> 2/n

SDFFit

2 (k 1) n

1 r232

Square of part correlation =


unique contribution of the
2
explanatory variable to R
(coefficient of multiple
determination).

=Hy

> 3 (3 sigma level)

r12 r13r23

(Hat Value)

Potential Outliers

Standardized
residual,
Studentized
residual

(1 r132 )(1 r232 )

21. Leverage Value

18. Outliers
Measure

r12 r13r23

Di =

j (Yj Yj(i))2/k x MSE


A (1-) 100% CI (Confidence
Interval) for the conditional
mean of Y given values of Xi:

r12,3

22. Relationship
between F and R2

PARTIAL CORRELATION
Correlation between y1 and x2,
when the influence of x3 is
removed from both y1 and x2.

Hat Matrix

You might also like