Digbym
Your two questions in this E-mail and also the other one in another E-mail concern with the same problem. A few scholars would like to refer to kriging as "multivariate linear regression". Indeed, kriging and least square resemble one another in all aspects except one. In kriging, residual is defined as estimated value minus true value (which is not accessible) while in least square, residual is defined as estimated value minus observed value. The notion of true value inaccessibility in kriging results in key concept including ergodicity, stationarity and .....
Have you ever asked yourself why we minimize sum of residuals squared and NOT sum of residuals cubed? The answer to this question concerns with normality of your data. Starting with Maximum Likelihood Method (MLM), one can easily link mean square error (MSE) to normality of data. In a few textbooks, it reads: "If your original data are normally distributed, then the chance to get better cross validation statistics will increase"
First few chapters of "Multivariate Geostatistics" by Wackernagel could be consulted for further detail.
Thanks Abedini On Sat, 2 Sep 2006, [EMAIL PROTECTED] wrote:
Why must data be transformed to normal for kriging, is it that the mean of mean of transformed data is a better estimate of the true mean, i.e. why is log estimate of mean a better estimate of the mean than the raw mean of logarithmic data? + + To post a message to the list, send it to [email protected] + To unsubscribe, send email to majordomo@ jrc.it with no subject and "unsubscribe ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
+ + To post a message to the list, send it to [email protected] + To unsubscribe, send email to majordomo@ jrc.it with no subject and "unsubscribe ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/
