DGCFit: Fit discrete multivariate data using Gaussian copula

In fanne-stat/DGCPCA: Discrete Gaussian Copula PCA

Description

Fit discrete multivariate data using Gaussian copula

Usage

DGCFit(mapped_data, Fhats, nearPDproj = T, ncores = 2, ...)

Arguments

mapped_data	A discrete dataset whose discrete values are all numeric or have been mapped to numbers.
nearPDproj	Logical. If TRUE, the estimated correlation matrix using Kendall's tau will be projected to the nearest positive definite matrix.
ncores	Number of cores to use when parallelized computing of the correlation matrix.
Fhat	A list of estimated marginal cdfs of the dataset. Each item in the list is a stepfun object and the support of the cdfs (knots(Fhats[[j]])) should be the superset of the unique discrete values of the corresponding feature (unique(mapped_data[,j])). If not specified, ecdf will be used to estimate the marginal cdfs.

Value

An object of class DGCFit with components

mapped_data	The input mapped_data.
tau	The Kendall's tau matrix of the input discrete numerical dataset.
corr	The estimated correlation matrix using the Kendall's tau matrix.
cdfs	The list of estimated marginal cdfs.

fanne-stat/DGCPCA documentation built on Dec. 20, 2021, 7:43 a.m.