arg.max: Find an argument that maximizes a function

In 42n4/dmr.util: A package with utility functions used in the book Cichosz, P. (2015): Data Mining Algorithms: Explained Using R

Description

This function applies a numeric-valued function to a vector or list of arguments and returns a specified number of arguments that yield the greatest function values. It is used in examples presented in the book Cichosz, P. (2015): Data Mining Algorithms: Explained Using R. See Appendix B or http://www.wiley.com/go/data_mining_algorithms for more details.

Usage

arg.max(args, fun, k = 1)

Arguments

args	a vector or list of function arguments
fun	a function that can be called with one argument and returns a numeric value
k	an integer specifying the number of arguments corresponding to the greatest function values to return

Details

The fun function is applied to each argument in args and then the k arguments that produced the greatest function values are returned. Ties are broken according to the original ordering of arguments (using stable sorting to determine the least k values).

Value

If args is a list, its sublist of length k containing the k arguments that yield the greatest fun values. If args is a vector, its subvector of length k containing the k arguments that yield the greatest fun values.

Author(s)

Pawel Cichosz <[e-mail]>

See Also

arg.min

Examples

arg.max(1:10, sin)
arg.max(1:10, sin, 3)

data(weatherr, package="dmr.data")
arg.max(1:nrow(weatherr), function(i) weatherr$playability[i], 3)
arg.max(weatherr$temperature, function(temp) abs(temp-20), 3)

42n4/dmr.util documentation built on May 20, 2019, 2:02 p.m.