Origin Version (Select Help-->About Origin): 7 Operating System: XP
When I do statistical weighting for a fit: does it mean that the higher my y-value, the more it is taken into consideration into the fit? i have exp. decaying data and i want that it puts more weight into the early (large) y-values as they have less error.
I presume you are referring to the weighting methods in NLSF tool. These are documented in the Help file in the section: Before you Start: The Chi-Square Minimization
As you can see from that help section, when you pick statistical weighting the weight is computed as (1/y-value), thus larger the y value less the weight.
You can pick other methods, including direct, where you can use a custom dataset for weight where you specify what value to use for each data point.
stat. weighting
Posted - 08/28/2005 : 2:51:17 PM
