Beruflich Dokumente
Kultur Dokumente
Y=B0+B1X1+B2X2+B3X3+B4X4+€
Step4: Assumptions
i. Linearity, Y, X1,X2,X3 and X4 are linearly related
ii. E(€ )=0
iii. € are uncorrelated
iv. € follows normal distribution
v. Variance of € = sigma^2 is constant(assumption of homoscedasticity)
vi. All regressor variables are independent
Step5:Final model under the assumptions
E(Y|X)= Y=B0^+B1^X1+B2^X2+B3^X3+B4X4
Step 6: Using the method of least squares the model will be built
Adjusted Rsquare from the excel sheet, (highly significant variation, more than
50%)
Step7: Model built is
Y^=B0+B1Si+B2Ex+B3Ru+B4So , where B0,B1,B2,B3,B4 are obtained from b
row of excel sheet(first row)
Step8: Interpretation of B coefficients
From the equation, define B0^, B1^, B2^, B3^, B4^.
B0^: when all the regressor variables are negligible level, the average value of commercial
effectiveness is at a level of -36.494
B1^: For every unit change in assessed sincerity, there will be 9.825 unit of change in commercial
effectiveness. Where all other regresson variable are at constant
B2^:
H0:B1=0
H1:B1≠0
Since P value for B1 is greater than α, B1 is insignificant
H0:B2=0
H1:B2≠0
Since P value for B2 is lesser than α, B2 is significant and it continues……….
B0, B2 are retained as they are significant
Step11: Rebuilding the model (B2 are retained as they are significant, P value is lesser than alpha)
Y^=B0+B2^X2
i. Linearity: high value of Rsquare = 0.9183 , and most of the scattered plot is close to
straight line hence the linearity is satisfied
ii. E(€)=0, since the average of error in (residual page of excel sheet) is very negligible this
is proved
iii. € are uncorrelated
D=
N=
K=1
dL= (table) -> 4-dL
dU=(table) -> 4-dU
dU<d<4-dU
Hence € are uncorrelated