competingPoints: select the most probable points based on a statistical model

In zarquon42b/RvtkStatismo: Integrating statismo and R using the vtkStandardMeshRepresenter

Description

select the most probable points based on a statistical model, using the mahalanobisdistance

Usage

competingPoints(model, sample, index, mahalanobis = TRUE)

## S4 method for signature 'pPCA,matrix,numeric'
competingPoints(model, sample, index,
  mahalanobis = TRUE)

Arguments

model	statistical shape model of class 'pPCA'
sample	k x 3 matrix of coordinates
index	integer vector of lenght k assigning the corresponding index of the model's coordinates to each row of sample
mahalanobis	logical: if FALSE, Euclideandistance is used.

Value

mahadistance	vector containing the mahalanobisdistances of all tested coordinates
goodverts	the coordinates with the lowest mahalanobisdistance
goodrows	integer vector containing the rows of sample that are selected
mahagood	mahalanobisdistances of the probable coordinates only

Examples

require(Morpho)
     data(boneData)
    align <- rigidAlign(boneLM)$rotated
    mymod <- statismoBuildModel(align,representer=align[,,1],sigma=2,scale=TRUE)
#add some arbitrary data
myconfused <- matrix(rnorm(300),100,3)
myconfusedind <- sample(1:10,size=100,replace=TRUE)
perturb <- sample(1:110)
out <- competingPoints(mymod,rbind(align[,,1],myconfused)[perturb,],c(1:10,myconfusedind)[perturb])
##check if the selected coords are identical to the actual ones
all.equal(align[,,1],out$goodverts)

zarquon42b/RvtkStatismo documentation built on May 4, 2019, 9:09 p.m.