impMean: Imputation of missing values in a collection of matrices

In phylter: Detect and Remove Outliers in Phylogenomics Datasets

Imputation of missing values in a collection of matrices

Description

Impute missing data in a list of matrices. Matrices are first given the same dimension, then missing entries are filled by computing the average value in non-missing corresponding entries in all matrices.

Usage

impMean(matrices)

Arguments

matrices

A list of distance matrices.

Value

Returns a list of matrices with same dimensions, with rows and columns in the same order and missing data (if any) imputed.

Examples

data(carnivora)
matrices <- phylter(carnivora, InitialOnly = TRUE, parallel = FALSE)$matrices

# remove n species randomly (n between 1 and 5) in each matrix (to mimic missing data)
fun<-function(mat) {
 species2remove<-sample(1:nrow(mat),sample(1:5,1))
 mat<-mat[-species2remove,-species2remove]
 return(mat)
}
matrices.missing<-lapply(matrices, fun)
#check that all matrices have now different dimensions: 
lapply(matrices.missing, dim)
# Impute data to get back to the same dimensions
matrices.ok<-impMean(matrices.missing)
lapply(matrices.ok, dim) #all dimensions are now identical. Missing data have been imputed. 

phylter documentation built on Aug. 24, 2023, 9:10 a.m.