View Single Post
  #1  
Old June 1st, 2010, 02:35 PM posted to microsoft.public.excel.worksheet.functions
PJ[_6_]
external usenet poster
 
Posts: 6
Default Regression Leverage Formula (Jerry W. Lewis or Mike Middleton)already have DFITS formula

I have been trying to figure out the leverage formula for a set of X Y
values
X Y
75 16
83 20
85 25
85 27
92 32
97 48
99 48

Using the data analysis add-in and adding residuals then running
regression gives me Predicted Y, Residuals and Standard Residuals, If
my data was in A2:B8 (X and Y) and C2 housed this formula =SLOPE($B
$2:$B$8,$A$2:$A$8)*A2+INTERCEPT($B$2:$B$8,$A$2:$A$ 8) which gives me my
predicted y and D2 housed B2-C2 which gives me my residuals then E2
housed =D2*D2 which gives me my residuals squared. What I am looking
for in cell F2 is HI1(the weight of the leverage value (y value) for
the ith residual). I have the values for the Leverage but do not know
how they were derived.

F2:F8
0.53588
0.200997
0.163787
0.163787
0.180066
0.331229
0.424252

I have the DFITS formula in G2:
=D2*SQRT((7-2-1)/(SUM($E$2:$E$8)*(1-F2)-E2))*SQRT(F2/(1-F2))

But do not know how they derived the Leverage Formula?
Can you please give me some guidance?
Thanks in advance