lmHyperframe: linear model of Kfunctions

In robertbagchi/ReplicatedPointPatterns: Analysis of replicated point patterns with covariates

Description

Function to take a hyperframe and fit a linear model to the data using all rows and the distance range Still very much in development so use carefully Not yet generalised to bivariate designs Not tested much yet

Usage

lmHyperframe(hyper, r, form, correction = "border", weights.type=NULL,
minsamp = NA, computeK=TRUE,  printwarnings = TRUE)
kfunclm(k, dat, form, weights)

## S3 method for class 'kfunctionlm'
print(x, ...)

Arguments

hyper	An object of class hyperframe (spatstat) containing point patterns and covariates
r	distances at which to model - must include only increasing positive numbers starting from 0
form	The names of the explanatory covariates in quotes
correction	The edge correction by name - options are those for Kest in spatstat. Defaults to 'border'
weights.type	The type of regression weights to be calculated. Options are 'nx', 'nx_A', 'nx2', 'nx2_A', 'sqrtnxny', 'nxny', 'nxny_A', 'sqrtnxny_A'
minsamp	Defines the minimum number of points that need to be in the central part of the plot (i.e. not affected by edge corrections). Defaults to NA so no cases are removed.
computeK	Controls whether hyperframe should compute K or if this is done externally.
printwarnings	Whether or not to print warnings
k	Vector of K(r) for all cases for a given distance r.
dat	Data set with covariates included in formula.
weights	Vector of weights for all cases for a given distance.
x	Object to be printed

Author(s)

Robert Bagchi Maintainer: Robert Bagchi <[e-mail]>

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (hyper, r, form, correction = "border", weights.type, 
    minsamp = 1, printwarnings = TRUE) 
{
    if (min(r) > 0) 
        r <- c(0, r)
    if (!all(c("pppx") %in% names(hyper))) {
        stop("hyperframe object must include 'pppx' element")
    }
    hyper$K <- with.hyperframe(hyper, Kest(pppx, r = r, correction = correction, 
        ratio = TRUE))
    hyper$wts <- with.hyperframe(hyper, list(kfunc.weights.calc(pppx, 
        r = K$r, correction = correction, type = weights.type)))
    sp.keep <- sapply(hyper$wts, function(x) all(unlist(x[r]) >= 
        minsamp))
    removed.species <- row.names(hyper)[!sp.keep]
    warning(paste("Removed", length(removed.species), "species with insufficient numbers"))
    hyper <- hyper[sp.keep, ]
    dist.keep <- (apply(sapply(hyper$K, function(K) K[[correction]]), 
        1, function(x) var(x)) > 0)
    warning(paste("Not modelling K at distances ", paste(r[!dist.keep], 
        collapse = ", "), "due to zero variance"))
    modr <- match(r[dist.keep], r)
    kmods <- sapply(modr, function(i) {
        kfunclm(k = sapply(hyper$K, function(k) k[[correction]][i]), 
            dat = as.data.frame(hyper, warn = FALSE), form = form, 
            weights = sapply(hyper$wts, function(w) unlist(w)[i]))
    }, simplify = FALSE)
    names(kmods) <- r[dist.keep]
    class(kmods) <- "kfunctionlm"
    return(kmods)
  }

robertbagchi/ReplicatedPointPatterns documentation built on May 27, 2019, 10:32 a.m.