[R-sig-Geo] Co-Kriging using two variables versus one variable

[Kai Zhang]

Hi folks,

I am working on an exercise which compares the performance of different Kriging methods including ordinary Kriging, Co-Kriging and Universal Kriging. Regarding Co-Kriging, I am trying to conduct Co-Kriging using two secondary variables separately. RMSEs from these two Co-Kriging are smaller than those from Ordinary Kriging. I am also conducting Co-Kriging using both secondary variables. However, RMSEs are larger than those from Co-Kriging with either secondary variable separately. I am wondering if these results do make sense since I expect that RMSEs from Co-Kriging using both variables are smaller. Any comments are welcome.

Thanks,
Kai 



__________________________________________________
赶快注册雅虎超大容量免费邮箱?