About V_chisq user defined curved fitting

Share icon

Tue, 02/09/2016 - 10:25 am


How do you specify the weighting wave mentioned below and how to derive a meaningful ChiSq value?


V_CHISQ: A measure of the goodness of fit. It has absolute meaning only if you've specified a weighting wave containing the reciprocal of the standard error for each data point.
The weighting wave should have one point for each of your fitted data points. The values should be your estimate of the standard error for the data points.

There is no particular advice on how to determine your estimate of the standard error of your data points. Some people are able to come up with a meaningful estimate based on a priori arguments, some use a sliding window and compute a standard deviation for the data within the window.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

February 10, 2016 at 09:46 am - Permalink

Use /W to specify the wave, and the /I flag if you have standard deviations in your weighting wave. The use of 1/SD and the need for /I is an unfortunate historical artifact.

John Weeks
WaveMetrics, Inc.
support@wavemetrics.com

February 10, 2016 at 05:06 pm - Permalink