Here’s a slightly different question: It is standard practice to test for significant differences in the parameter coefficient estimates to determine “difference” between alternative models?

Thus, if one or more of the coefficients are significantly different, then the models are deemed different.

This question arises from wanting to test for changes in spatial and/or temporal patterns in image time series.

Cheers, Geoff

|"|"|"|""//*\\""|"|"|"|

Geoffrey M. Henebry, Ph.D., C.S.E.

Collaboratory for Ecological Remote Sensing, Modeling, and Analysis (CERSMA)

Center for Advanced Land Management Information Technologies (CALMIT)

School of Natural Resources

University of Nebraska-Lincoln

102 E. Nebraska Hall

Lincoln, NE 68588-0517 USA

voice: 402-472-6158 (-4608 fax)

email: ghenebry@...

http://calmit.unl.edu/cersma/

**From:**Digby Millikan [mailto:digbym@...]

**Sent:**Wednesday, April 20, 2005 2:48 AM

**To:**ai-geostats; Lee Sung Soo

**Subject:**Re: [ai-geostats] How can I measure the difference between 2 models?Is you model 2 or 3 dimensional what about a map of the errors? Also general statisitics

of the populations, e.g. varaiance, coefficient of variation etc. Grade tonnage plot is a good

idea.

Regards Digby

----- Original Message -----

**From:**Lee Sung Soo**To:**ai-geostats@...**Sent:**Thursday, April 14, 2005 4:40 PM**Subject:**[ai-geostats] How can I measure the difference between 2 models?Hello everyone,

I apologize if this question is too elementary for this list;

The problem is how to quantify and represent the difference between

standard model and simulated model. I tried to use a tonnage and

grade mean value difference by cut-off grade in order to compare

and quantify some risks between models.

But I think there are some limitations to express the variations related to the grade

distribution, although this analysis is popular and common.

For instance, please see below figure.

Standard model(A)

************

* 55 90 *

* *

* *

* 10 45 *

************

Simulated model(B)

************

* 90 45 *

* *

* *

* 55 10 *

************

The mean values of two block models show exact same and look very

similar. But the grade distribution is pretty different.

To express the difference and variabilities, some more various ways

are required. How can I measure the difference between 2 models.

Would you give me a advice or ways about quantifying the difference

between two models. And if available, would you introduce me

reference papers or internet site.

Thanks

Best regards,

Lee Sung Soo

Master student(University of WA)

* By using the ai-geostats mailing list you agree to follow its rules

( see http://www.ai-geostats.org/help_ai-geostats.htm )

* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to sympa@...

Signoff ai-geostats- Colin Badenhorst was discussing using F and T - test's to compare statisitics of two model's, as discussed in most graduate level statistics text's e.g. Miller & Freund's Probability and Statistics for Engineers etc.----- Original Message -----
**From:**Din Chen**Sent:**Thursday, April 21, 2005 1:00 AM**Subject:**RE: [ai-geostats] How can I measure the difference between 2 models?Dan Griffith in his book (the Introduction) Advanced Spatial Statistics explained exactly how to deal with this problem. You can not use the mean since mean is not a sufficient statistics and also it is easy to prove that the mean is not an efficient statistics.

**From:**Geoff Henebry [mailto:ghenebry@...]**Sent:**Wednesday, April 20, 2005 6:43 AM**To:**ai-geostats@...**Subject:**RE: [ai-geostats] How can I measure the difference between 2 models?Heres a slightly different question: It is standard practice to test for significant differences in the parameter coefficient estimates to determine difference between alternative models?

Thus, if one or more of the coefficients are significantly different, then the models are deemed different.

This question arises from wanting to test for changes in spatial and/or temporal patterns in image time series.

Cheers, Geoff

|"|"|"|""//*\\""|"|"|"|

Geoffrey M. Henebry, Ph.D., C.S.E.

Collaboratory for Ecological Remote Sensing, Modeling, and Analysis (CERSMA)

Center for Advanced Land Management Information Technologies (CALMIT)

School of Natural Resources

University of Nebraska-Lincoln

102 E. Nebraska Hall

Lincoln , NE 68588-0517 USA

voice: 402-472-6158 (-4608 fax)

email: ghenebry@...

http://calmit.unl.edu/cersma/

**From:**Digby Millikan [mailto:digbym@...]**Sent:**Wednesday, April 20, 2005 2:48 AM**To:**ai-geostats; Lee Sung Soo**Subject:**Re: [ai-geostats] How can I measure the difference between 2 models?Is you model 2 or 3 dimensional what about a map of the errors? Also general statisitics

of the populations, e.g. varaiance, coefficient of variation etc. Grade tonnage plot is a good

idea.

Regards Digby

----- Original Message -----

**From:**Lee Sung Soo**To:**ai-geostats@...**Sent:**Thursday, April 14, 2005 4:40 PM**Subject:**[ai-geostats] How can I measure the difference between 2 models?Hello everyone,

I apologize if this question is too elementary for this list;

The problem is how to quantify and represent the difference between

standard model and simulated model. I tried to use a tonnage and

grade mean value difference by cut-off grade in order to compare

and quantify some risks between models.

But I think there are some limitations to express the variations related to the grade

distribution, although this analysis is popular and common.

For instance, please see below figure.

Standard model(A)

************

* 55 90 *

* *

* *

* 10 45 *

************

Simulated model(B)

************

* 90 45 *

* *

* *

* 55 10 *

************

The mean values of two block models show exact same and look very

similar. But the grade distribution is pretty different.

To express the difference and variabilities, some more various ways

are required. How can I measure the difference between 2 models.

Would you give me a advice or ways about quantifying the difference

between two models. And if available, would you introduce me

reference papers or internet site.

Thanks

Best regards,

Lee Sung Soo

Master student( University of WA )

* By using the ai-geostats mailing list you agree to follow its rules

( see http://www.ai-geostats.org/help_ai-geostats.htm )

* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to sympa@...

Signoff ai-geostats

* By using the ai-geostats mailing list you agree to follow its rules

( see http://www.ai-geostats.org/help_ai-geostats.htm )

* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to sympa@...

Signoff ai-geostats