findCombination: Find a dimension/combination of dimensions with certain...

In rwoldford/tidy-table: Tidy Up And Rearrange Tables To Better Reveal Data Patterns

Description

Find a dimension/combination of dimensions with certain conditions.

Usage

findCombination(table, combs = "NULL", chooseFun = "median",
  valueFun = "mad", dimDecreasing = TRUE)

Arguments

table	A sample of data with dimensions >=2, can be a matrix, an array or a table.
combs	A matrix of dimension combinations, each column stores the dimensions of a combination.
chooseFun	The summary to represent the variability (spread) of each dimension/combination of dimensions, can be functions in R or self-defined functions.
valueFun	The summary to measure the variability (spread) of each layer of the dimension/combination of dimensions, can be functions in R or self-defined functions.
dimDecreasing	Logical, if order the dimensions/combinations in a decreasing order.

Value

The dimension/combination of dimensions which is ordered first.

Note

Here, the default of chooseFun is "median" and the default of valueFun is "mad" since in our tidytable package we define the varaibility of a dimension (combination of dimensions) as the median of mad values get from each layer of that dimension (combination).

Author(s)

R. Wayne Oldford and Xiaomei Yu

Examples

table <- Titanic
coms <- combn(length(dim(table)), 2) # all the combinations of two dimensions
findCombination(table, coms, chooseFun = "median", valueFun = "mad", dimDecreasing = TRUE)
# so the combination of first two dimensions has the largest variability
# among all combinations of two dimensions in Titanic data

rwoldford/tidy-table documentation built on May 12, 2019, 4:38 a.m.