[R-sig-Geo] Statistical inference about spatially varying coefficients of a mixed GWR model

[Dongwoo Kang]

Dear R-sig-geo list,

Hi everyone, can I ask your help about statistical inference of a mixed GWR
model?
Thank you so much in advance.

I am working on my PhD dissertation and it is about the spatial
heterogeneity of regional knowledge producing based on the data of 3,109 US
counties.

My response variable is "the number of patents per capita" and I am
controlling 11 explanatory variables including constant term.

By using the Monte Carlo simulation function "montecarlo.gwr" in "GWmodel"
of R package, 
I identify 4 global variables and 7 spatially varying variables.

Following Mei, Wang, and Zhang (2006), I could calculate the p-values for 4
global variables by using a 
bootstrap method. In other words, I run an unrestricted MGWR as an
alternative hypothesis and compare its RSS with the RSS of restricted MGWR
for each global variables as an null hypothesis of beta=0.

But my question is how to calculate the p-values for each local estimates of
spatially varying coefficients.
In my case, I have 7 spatially varying variables and 3,109 observations,
thus I want to get the p-values of 21,763 local estimates.

In addition, when I run a basic GWR model by using "gwr.basic" function of
"GWmodel" package,
"gwr.t.adjust" function allows us to get adjusted p-values for each local
spatially varying coefficients by using Bonferroni type standard error
correction.

Ultimately, I want to calculate Bonferroni correction standard errors and
adjusted p-values for each spatially varying local coefficients in my mixed
GWR model.

If you have any idea about my questions, please help me.

Thank you!

With best regards,
Dongwoo Kang


References
Mei, C. L., N. Wang and W. X. Zhang (2006). "Testing the importance of the
explanatory variables in a mixed geographically weighted regression model."
Environment and Planning A 38(3): 587-598.





--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/Statistical-inference-about-spatially-varying-coefficients-of-a-mixed-GWR-model-tp7587270.html
Sent from the R-sig-geo mailing list archive at Nabble.com.