How to predict the value for a location using other predictor variables?

[fabefabe]

Hi all
I have a point shapefile that has 3 variables like precipitation, temperature, and the runoff. I am using precipitation and temperature as predictor variables and runoff as dependent variable. Using the spatial relation of these variables, I need to predict runoff for the points in a shapefile for any changes in temperature and precipitation as shown below.

Eg
ID PPT TEMP RUNOFF
1 110 25 40
2 100 18 34

ID PPT TEMP RUNOFF
1 150 24 ?
2 90 17 ?

I think this problem is different then interpolation because I am not interested in predicting for unknown locations as well as I have multiple variables to be taken care of. I am just wondering if I can perfom any test like this in ArcGIS? If so, which tool should I use? Any light on this area would be highly appreciated?

Thanks!

[JeffreyEvans]

Given the problem that you stated I do not know what you mean by "spatial relationships". It sounds like a simple OLS or GWR (both available in ArcGIS) would work. If for some reason these approaches will not work consider imputation. For an imputation analysis you will need to use a statistical software. R has very nice capacity to read/write GIS data and very mature imputation approaches (and is free). Take a look at this thread.

http://forums.arcgis.com/threads/25949-cluster-analysis-based-on-multiple-variables?highlight=imputa...

[LaurenRosenshein]

Hi Kami,

Examining the relationships between the variables that you have in your data can be done using Ordinary Least Squares regression and Geographically Weighted Regression, both available in ArcGIS.  To get started using OLS and GWR, you may want to check out the Regression Analysis Basics seminar, which is free from the Virtual Campus website.

You may also want to check out the other resources available for spatial statistics here:  http://esriurl.com/spatialstats.

Hope this helps.

Lauren Rosenshein
Geoprocessing Product Engineer