|
|
Hi, I'm new of here so please be patient.
I am trying to determine the coefficient of correlation (R2) to a theoretical gaussian distribution of some actual distributions I got from subjects.
I use the following equation model (in the script)
*************************************************************************
* NonLinear Regression.
MODEL PROGRAM Amplitude=1 Mean=1 SD=0.1.
COMPUTE PRED_=Amplitude * EXP( - 0.5 * ((X - Mean) / SD) ** 2).
NLR y1
/OUTFILE='C:\Users\Lorenzo\AppData\Local\Temp\spss6284\SPSSFNLR.TMP'
/PRED PRED_
/CRITERIA SSCONVERGENCE 1E-8 PCON 1E-8.
*************************************************************************
My actual data are on a 40 1-second time bins (X axes 1 to 40) and my frequencies have been normalized so that the peak = 1.0. The theoretical position of the peak should be at the 20th second.
The ANOVA output says:
***************************************************************************
Dependent variable: y1
a. The solution results in negative sum of squares. Check
model specification.
b. R squared = 1 - (Residual Sum of Squares) / (Corrected
Sum of Squares) = ..
***************************************************************************
And doesn't give any R2 value
Does anybody know what I am doing wrong ?
Thank-you in advance. Lorenzo
|
Locked
