| Author |
 Topic  |
|
|
gabrielscolon
Mexico
Posts |
 Posted - 09/09/2005 : 12:30:19 AM
|
Origin Version (Select Help-->About Origin):
Operating System:W2000
is it possible to estimate parameters of a set of functions (dependent variables yi, data and error for each yi) with no independent variables xi?
|
|
|
minimax
357 Posts |
Posted - 09/09/2005 : 05:03:17 AM
|
Hi gabrielscolon,
What you really mean is the autoregression model, such as Y(t)=+*Y(t-1)+(t)? If so, we are planning to develop a time series module in the future version of Origin.
Max
OriginLab GZoffice |
 |
|
|
gabrielscolon
Mexico
Posts |
Posted - 09/09/2005 : 2:00:11 PM
|
Hi minimax,
no it is not. Maybe is just my ignorance, what I mean is a routine like CERN's minuit. I want to use Origin to estimate the parameters, lets say, p1 and p2, that minimize the chi-squared function:
chisq(p1,p2) = ((y1-f1(p1,p2))/err1)^2 + ((y2-f2(p1,p2))/err2)^2
+ ((y3-f3(p1,p2))/err3)^2,
the input are the theoretical expressions for f1, f2, and f3, in terms of p1 and p2, and their corresponding experimental values y1, y2, and y3, along with their standard deviations err1, err2, and err3.
Thank you for your help.
gabriel
|
|
|
| |
Topic |
|
|
|
fitting with no independent variables