calculate_dist: Calculate Hellinger's Squared Distance
In ZytoHMGU/hetset: Identification of Heterogeneous Subsets in Data

Description Usage Arguments Details Value Author(s) References See Also Examples

Hellinger's squared distance is calculated for the subpopulations of an hetset object.

1	calculate_dist(H)

`H`	`hetset` object with metadata about parameters as created by `estimate_densities`

A hetset object is a SummarizedExperiment container with information about a two-component normal mixture model fitted to a subset of features of the first assay of the container.

Once, parameters of the component-densities are listed in the metadata (e.g. by estimate_densities or scan_hetset), Hellinger's squared distance of the two densities can be calculated with this function.

For inversion of the covariance matrix, the matlib package is used.

The same hetset object is returned with an updated sqHell-entry in the metadata.

Daniel Samaga

Pardo, L. (2006). Statistical Inference Based on Divergence Measures. New York: Chapman and Hall/CRC. page 51.

Michael Friendly, John Fox and Phil Chalmers (2018). matlib: Matrix Functions for Teaching and Learning Linear Algebra and Multivariate Statistics. R package version 0.9.1. https://CRAN.R-project.org/package=matlib

hetset, scan_hetset, estimate_densities

A <- matrix(data = rnorm(n = 2000,mean = 1,sd = 1),ncol = 20)
B <- matrix(data = rnorm(n = 2000,mean = 5,sd = 1),ncol = 20)
Hds <- hetset(D = cbind(A,B))
rm(A,B)
Hds$prt <- as.factor(sample(c("A","B"),ncol(Hds),TRUE))
Hds@metadata$slf <- c("F5","F15")
Hds <- estimate_densities(Hds)
Hds <- calculate_dist(Hds)
print(Hds@metadata$sqHell)