Hi,
SSErr is known to be a fit statistic to determine how well the variogram
model fit the data or sample variogram. Please I would like to know how
this statistic (SSErr) is being calculated. With gstat, is it possible to
use other fit statistics such as AIC, BIC?
In my own case, SSErr identifies exponential to be the best that fit the
data but the cross validation statistics (RMSE, ME, MSDR) identifies
Spherical and then exponential to be optimal for prediction accuracy. Is
this right? My thought is that if exponential is best fitting model as
identified by SSErr, it should be the best in terms of prediction accuracy
by cross validation. Please comment and educate me on this issue.
Thank you.
Moshood
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
