meanAggregateClusterProbability {isocat}

Create mean aggregate probability-of-origin surfaces for each cluster.

Description

Subset probability-of-origin surfaces by cluster assignment and find mean aggregate probability-of-origin surface for each clustered group.

Usage

meanAggregateClusterProbability(indivIDs, clusters, surfaces, nClust = FALSE)

Arguments

indivIDs

Vector of individual ID variables corresponding to surface names.

clusters

Vector of cluster IDs, in an order corresponding to 'indivIDs'.

surfaces

Stack of probability-of-origin surfaces for all individuals. Object of class 'RasterStack.'

nClust

Create and apply a multi-core cluster for faster processing using 'raster' and 'parallel' packages. Defaults to 'FALSE' (i.e., no clustering).

Examples

# Create and cluster example assignment surfaces.
myiso <- rasterFromXYZ(isoscape)
myiso_sd <- rasterFromXYZ(isoscape_sd)
df <- data.frame(
        ID = LETTERS[1:9],
        isotopeValue = seq(-120,-25,length.out = 9),
        SD_indv = rep(5, 9)
        )
assignmentModels <- isotopeAssignmentModel(
         ID = df$ID,
         isotopeValue = df$isotopeValue,
         SD_indv = df$SD_indv,
         precip_raster = myiso,
         precip_SD_raster = myiso_sd,
         nClusters = FALSE
         )
mySimilarityMatrix <- simmatrixMaker(assignmentModels)
cS <- clusterSimmatrix(
         simmatrix = mySimilarityMatrix,
         r = seq(.7,1.4,by=.1)
         )
# Cut clusters.
myheight <- 0.25
df$cluster <- dendextend::cutree(cS$hclust, h = myheight)
# Create mean aggregate surfaces.r p
meanSurfaces <- meanAggregateClusterProbability(
         indivIDs = df$ID,
         clusters = df$cluster,
         surfaces = assignmentModels,
         nClust = FALSE
         )