Randomly impose a monotone missingness pattern by replacing the ends
of each column of the input matrix by a random number of NAs
Usage
rmono(x, m = 7, ab = NULL)
Arguments
x
data matrix
m
minimum number of non-NA entries in each column
ab
a two-vector of alpha (ab[1]) and
beta (ab[2]) parameters to a
Beta(alpha, beta) distribution
describing the proportion of NA entries in each column.
The default setting ab = NULL yields a uniform distribution
Details
The returned x always has one (randomly selected)
complete column, and no column has fewer than m
non-missing entries. Otherwise, the proportion of missing entries
in each column can be uniform, or it can have a beta
distribution with parameters alpha (ab[1]) and
beta (ab[2])
Value
returns a matrix with the same dimensions as the input x