MISS: Creating Missing Values at Random in Multivariate Datasets

In PAUL1993-P/CondMVT.package: Conditional Multivariate t Distribution, Expectation Maximization Algorithm, and Its Stochastic Variants

Creating Missing Values at Random in Multivariate Datasets

Description

This function randomly creates missing values in a multivariate dataset. The resultant missing data mechanism is missing at random (MAR). The percentage of missingness has to be specified. This percentage is computed as a proportion of the sample size. In addition, the function allows for more than one missing value in any given case. It is set such that in a p-variate dataset, for any i^{th} case, the maximum allowable number of missing values is p-1. This helps avoid a situation where a case has no observed value.

Usage

MISS (TT, Percent)

Arguments

TT	n×p complete dataset.
Percent	the proportion of missing values, which must be specified.

Value

Data Y of size n×p with missing values (NA) created at random. The missing values are logical in nature.

Examples

# 3-dimensional multivariate t distribution
n <- 10
p=3
df=3
mu=c(1:3)
A <- matrix(rt(p^2,df), p, p)
A <- tcrossprod(A,A) #A %*% t(A)

Y7 <-mvtnorm::rmvt(n, delta=mu, sigma=A, df=df)
Y7
TT=Y7 #Complete Dataset

#Introduce MAR Data
Y8= MISS(TT,20) #The newly created incomplete dataset.
Y8  

PAUL1993-P/CondMVT.package documentation built on July 2, 2022, 1:49 a.m.