I conducted similar simulations using invIrM in the spdep package. Dr. Anselin has a nice tutorial entitled Spatial Regression Analysis in R, A Workbook ( http://www.sal.uiuc.edu/stuff/stuff-sum/pdf/rex1.pdf ) (90 pp., 456K) at http://www.sal.uiuc.edu/stuff/stuff-sum/tutorials where I began to learn the process. Terry
>>> Kitty Lee <[EMAIL PROTECTED]> 6/2/2007 2:36 PM >>> Yes, let me be more specific. Say I have two variables, X and Y. I know that at the individual level, the relationship between them is: Y=a + bX+ error I try to run some simulation where X is spatially clustered. I can generate spatially clustered points, e.g. nsp.u2<-pcp.sim(25,40,.001, sploy) where sploy is the boundary. But after generating the points, how can I assign values of X and Y, such that when I run moran I on X and Y, I will get a high spatial autocorrelation values? If I generate X and Y this way: X<-runif(nrow(nsp.u2),20,150) err<-rnorm(nrow(nsp.u2)) ratio<-.75 Y<-ratio*X+err And then do cbind(nsp.u2, X, Y). The X and Y in the dataset are not spatially clustered. Any ideas on how I can make X and Y spatially clustered? Thanks!! K. On Fri, 1 Jun 2007, Kitty Lee wrote: > Dear R-users, > > I'm trying to do some spatial simulation. I have two covariates, Z and > C. I want to examine their relationship under different spatial > distribution. > > I have no problem simulating completely spatial random process but I'm > totally stuck on poisson (cluster) pattern. I already have a dataset > with Z and C (obs=575) and I know the relationship between them. Using > these 575 cases, how can I simulate a clustered process and have a > dataset that has four columns: > > x-coor y-coor z c > > I know I can use rpois or pcp.sim to generate points that clustered and > then use cbind to attach Z and C values. But the problem is my > observations will not be spatially clustered. How can I simulate so that > now Z is spatially clustered? Although you are not being very clear, I think that unless both Z and C are marked point processes (ie. both take (discrete) values and are observed at different points), this is not a spatial point process problem. If Z and C are observed at the same points, and what you are looking for are clusters of correlated values of Z with C, the clusters are not the locations of the points, but rather the co-occurrence of high/low values of Z and C respectively. How to go forward from here would depend on what kind of process you are actually interested in. More detail might help - as might following up on R-sig-geo, rather than on the general list. > > Thanks! > > K. > > > --------------------------------- > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R ( http://www.r/ > )-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > -- Roger Bivand Economic Geography Section, Department of Economics, Norwegian School of Economics and Business Administration, Helleveien 30, N-5045 Bergen, Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43 e-mail: Roger.Bivand at nhh.no --------------------------------- Building a website is a piece of cake. [[alternative HTML version deleted]] _______________________________________________ R-sig-Geo mailing list R-sig-Geo@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo [[alternative HTML version deleted]] _______________________________________________ R-sig-Geo mailing list R-sig-Geo@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo
