This function creates a set of consistently rounded frequency count tables or hypercubes by means of a version of small count rounding.
makeroundtabs(
A,
b = 3,
d,
micro = "TRUE",
sort,
control,
nin = "n",
nout = "n",
minit = 3,
maxit = 3,
maxdiff = 5,
seed
)A data frame representing a micro dataset or a frequency count hypercube. The (first) columns define the variables. If A is a hypercube the last column contains the number of units in each cell. If A is a micro dataset it is reduced to hypercube by the function aggrtab.
Rounding base. Counts in A less than b tat are contributing to counts less than b in the marginal cubes D are selected from A. The selected dataframe is called B
A list d[[j]] whose elements are vectors of variable names from A defining marginal tables/cubes D of A that we are interested in.
Logical. TRUE if A is a micro dataset (default). FALSE if A i a frequency count hypercube.
An ordered list of variables in hypercubes in D meant for priority sorting of the reduced hypercube B before rounding. Not all variables in D should be included.
A list of marginals of the hypercubes in D where deviations of aggregated rounded counts are checked against original counts.
Name of count variable if A is a hypercube. Default name: "n".
Name of the frequency count variable in the output tables.
Minimum number of searches to be carried out.
Maximum number of searches to be carried out.
If maximum difference in "control" is no larger than maxit, the stop search.
Input seed for first systematic random search.
Ar: The rounded version of A
Br: The rounded version of B
D: The original hypercube of interest.
Dr: The rounded version of D. The final table of interest.
maxdiff: The largest absolute difference between cells D and Dr among cells in the control list.
nmaxdiff: The number of occurences if Maxdiff