nonlinear regression methods

Question

I've been exploring the Spatial Statistics tools in ArcGIS 10.1, specifically the Modeling Spatial Relationships toolbox.  From reading the help documentation, I've gathered that these tools are only appropriate for scenarios in which the relationships between explanatory and dependent variables are linear.  For scenarios with nonlinear relationships, the help documentation suggests either transforming the data to make it linear, or using a nonlinear regression method.  Does anyone have any advice about moving forward with either of these suggestions for working with nonlinear relationships?  I don't know how I would go about transforming my data, and I haven't been able to find any nonlinear regression tools in ArcGIS.

ChristopherBevilacqua · Answer

Thanks for the reply @Dan Patterson.  I'm looking at gas well production and several variables that influence it.  It turns out that the engineers I'm supporting may view the data relationships as linear.  I have a follow-up meeting with one of them next week to discuss.  I'll post an update afterward.

DanPatterson_Retired · Answer

Are you working with spatial data or essentially attribute data?

By that I mean if location is not the independent or dependent variable, then you don't need to deal with spatial statistics.  If you are interested in whether something is occurring/appearing at locations then transformation of the dependent variable is less of an issue as it would be in conventional linear regression between two variables.

Perhaps, you could tell us what you are looking at doing and the data you have to solicit some suggestions before you go down the transformation path needlessly (or with targeted direction)