Hello, all,

In linear regression, Y1 is dependent variable, Y2 is predicted value.
R is Pearson's r for Y1 and Y2, so I can get R square.
I can also get R square by the following formula.
R square = 1 - SSE/CSS
WHERE                                  
SSE = the sum of squares for error
CSS = CORRECTED TOTAL SUM OF SQUARES FOR THE DEPENDENT VARIABLE.

I found the two values are different. So I think I can only use the
second formula for nolinear regrssion, in linear regression, I can
only calculate pearson's r. Is it right?

thank you very much.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to