binTable: Sum all rows and all columns in a table, except for the...

In soundcorrs: Semi-Automatic Analysis of Sound Correspondences

Description

Useful for when the data are scarce and chisq.test returns a warning, or when a more specific analysis of the data is required.

Usage

binTable(x, row, col)

Arguments

x	[data.frame/matrix/table] Table to be binned.
row	[integer/vector] The rows to not be binned.
col	[integer/vector] The columns to not be binned.

Details

When working with sparse data, the absolute values in a table are sometimes too low to allow for the use of various statistical tests, or the features too numerous for the result of a statistical test to be clearly interpretable. In such cases, a solution may be found in binning, i.e. in combining all the rows or columns into one, with the exception of select few. For example, a 10x10 table may be thus reduced to a 2x2 or a 2x3 one. The values are magnified while the number of features is reduced.

Value

[table] Table with some of its data binned.

Examples

mtx <- matrix (1:16, nrow=4, dimnames=list(paste0("r",1:4),paste0("c",1:4)))
binTable (mtx, 1, 1)
binTable (mtx, 1, c(1,3))

soundcorrs documentation built on Nov. 16, 2020, 5:09 p.m.