Scribd
Upload
66 views

Final Cheat Sheet!

1. Single variable regression relates the mean of the dependent variable (y) to the independent variable (x) through a hypothesized deterministic model. Parameters are estimated using sample data. 2. Multiple regression also relates y to multiple x values. It makes assumptions about the distribution and independence of the random error term. Parameters are estimated to minimize unexplained error. 3. Statistical tests evaluate whether parameters are significantly different from zero and the overall model fit. Metrics like R-squared are used to assess model usefulness for estimation and prediction.

Uploaded by

target1010101
Copyright
© Attribution Non-Commercial (BY-NC)
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
66 views

Final Cheat Sheet!

1. Single variable regression relates the mean of the dependent variable (y) to the independent variable (x) through a hypothesized deterministic model. Parameters are estimated using sample data. 2. Multiple regression also relates y to multiple x values. It makes assumptions about the distribution and independence of the random error term. Parameters are estimated to minimize unexplained error. 3. Statistical tests evaluate whether parameters are significantly different from zero and the overall model fit. Metrics like R-squared are used to assess model usefulness for estimation and prediction.

Uploaded by

target1010101
Copyright
© Attribution Non-Commercial (BY-NC)
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/ 1

SINGLE VARIABLE REGRESSION 2

( x p− x́ )

√
1. Hypothesize deterministic components of^y model 1
± t ∝ s 1+ + DF n−2
relating mean E(y) to x. 2
n ( )
(∑ ) x
2

2. Use sample data to estimate unknown parameters. ∑ ( x 2 )− n


3. Specify Pop. Dist. Of random error (ɛ) and estimate
std. dist. of ɛ.
4. Statistically evaluate usefulness of model.
5. Use model for estimation, prediction, etc…
MULTIPLE REGRESSION
(ɛ) ASSUMPTIONS: random error
SSE
1. The mean of the prob. dist. of ɛ is 0. S2=
2. The variance of the prob. dist. of ɛ is constant for n−( k +1)
^β i ± t ∝ s ^ =C . I . for DF n−(k +1)
3.
all x values.
The prob. dist. of ɛ is 0. 2
β
( ) i

^β
4. Values of ɛ associated w/ any 2 values of y are
H o : ^β i=0∧H a : β^ i ≠0 t= i
independent. s ^β i
^y = β^ 0 { ý − ^β1 x́ }+ ^β 1 ¿ SECOEF :s β^
i

SSE(Unexplained Error)=∑ ( y 2 )− ^β 0 ∑ y− ^β 1 ∑ xy 2 SSR yy


SSE R=
S= Watch for High Leverage Pts. SST yy
n−2 n−1
s SRi >2 means outlier R a2 = (1−R 2)
s ^β = n−( k +1 )
1 2

√
∑x− n
^
2 (∑ )

^β 1 test :t= β 1−β 1


x
GLOBAL F
H o : β 1=β 2=β 3=β 4 =…=β i=0
H a : At least one of theabove coefficientsis nonzero .
s ^β 1 SSR R2
^β 1 ±(t ∝ )s ^ =(Confidence Interval for Slope DF n−2)
β1 k k
2 F= =
SSE 1−R2
∑ x∑ y n−(k +1) n−(k +1)
SS xy ∑ xy− n
r= =
√ SS xx SS yy 2 2
NESTED F

r 2=
√[
SSR(explained error )
=
2
∑ ( x )−
(∑ x )
n ][
∑ ( y )−
SST −SSE
2 (∑ y )
n ] H o : β g=β g+1=β g+2=…=β k =0
H a : At least one of theabove coefficie nts is nonzero.
SSER −SSEC
2
SST (∑ y )
2 k −g
∑ ( y )− n
F=
SSEC
ρ−test :t=r √ n−2 ( H : ρ=0 ) DF n−2 n−( k +1 )
2 0
√ 1−r v1 =k−g v 2=n−( k +1)
2
( x p − x́ )

√
1
^y ± t ∝ s + DF n−2
( )
2
n (∑ x )
2

∑ ( x 2 )− n

You might also like