spatialkernel-package     package:spatialkernel     R Documentation

_T_h_e _S_p_a_t_i_a_l_k_e_r_n_e_l _P_a_c_k_a_g_e

_D_e_s_c_r_i_p_t_i_o_n:

     An R package for spatial point process analysis.

_D_e_t_a_i_l_s:

     This package contains functions for spatial point process analysis
     using kernel smoothing methods. This package has been written to
     be compatible with the 'splancs' package which is available on
     CRAN (The Comprehensive R Archive Network).

     For a complete list of functions with individual help pages, use
     'library(help =  "spatialkernel")'.

_M_a_i_n_t_a_i_n_e_r:

     Pingping Zheng pingping.zheng@lancaster.ac.uk

_N_o_t_e:

     For the convience of the user, we present here examples which show
     how to use some of the functions in the package.

_A_u_t_h_o_r(_s):

     Pingping Zheng and Peter Diggle

_R_e_f_e_r_e_n_c_e_s:

        1.  P. Zheng, P.A. Durr and P.J. Diggle (2004) Edge-correction
           for Spatial Kernel Smoothing - When Is It Necessary?
           _Proceedings of the GisVet Conference 2004_, University of
           Guelph, Ontario, Canada, June 2004.

        2.  Diggle, P.J., Zheng, P. and Durr, P. A. (2005)
           Nonparametric estimation of spatial segregation in a
           multivariate point process: bovine tuberculosis in Cornwall,
           UK. _J. R.  Stat. Soc. C_, *54*, 3, 645-658.

_S_e_e _A_l_s_o:

     'cvloglk', 'phat', 'mcseg.test', 'plotphat', 'plotmc', 'pinpoly',
     'risk.colors', 'metre'

_E_x_a_m_p_l_e_s:

     ## An example of spatial segregation analysis
     ## Not run: 
     ## source in Lansing Woods tree data within a polygon boundary
     data(lansing)
     data(polyb)
     ## select data points within polygon
     ndx <- which(pinpoly(polyb, as.matrix(lansing[c("x", "y")])) > 0)
     pts <- as.matrix(lansing[c("x", "y")])[ndx,]
     marks <- lansing[["marks"]][ndx]
     ## select bandwidth
     h <- seq(0.02, 0.1, length=101)
     cv <- cvloglk(pts, marks, h=h)$cv
     hcv <- h[which.max(cv)]
     plot(h, cv, type="l")
     ## estimate type-specific probabilities and do segregation tests
     ## by one integrated function
     sp <- spseg(pts, marks, hcv, opt=3, ntest=1000, poly=polyb)
     ## plot estimated type-specific probability surfaces
     plotphat(sp)
     ## additional with pointwise significance contour lines
     plotmc(sp, quan=c(0.025, 0.975))
     ## p-value of the Monte Carlo segregation test
     cat("\np-value of the Monte Carlo segregation test", sp$pvalue)

     ##estimate intensity function at grid point for presentation
     ##with bandwidth hcv
     gridxy <- as.matrix(expand.grid(x=seq(0, 1, length=101), y=seq(0, 1, length=101)))
     ndx <- which(pinpoly(polyb, gridxy) > 0) ##inside point index
     lam <- matrix(NA, ncol=101, nrow=101)
     lam[ndx] <- lambdahat(pts, hcv, gpts = gridxy[ndx,], poly =
         polyb)$lambda
     brks <- pretty(range(lam, na.rm=TRUE), n=12)
     plot(0, 0, xlim=0:1, ylim=0:1, xlab="x", ylab="y", type="n")
     image(x=seq(0, 1, length=101), y=seq(0, 1, length=101),
         z=lam, add=TRUE, breaks=brks, col=risk.colors(length(brks)-1))
     polygon(polyb)
     metre(0, 0.01, 0.05, 0.51, lab=brks, col=risk.colors(length(brks)-1), cex=1)

     ## An example of inhomogeneous intensity function and K function
     ## estimated with the same data
     s <- seq(0, 0.06, length=101)
     lam <- lambdahat(pts, hcv, poly=polyb)$lambda
     kin <- kinhat(pts, lam, polyb, s)
     plot(kin$s, kin$k-pi*(kin$s)^2, xlab="s", ylab="k-pi*s^2", type="l")
     ## End(Not run)

