msu: Estimating Multivariate Symmetrical Uncertainty.

In msu: Multivariate Symmetric Uncertainty and Other Measurements

Description

MSU is a generalization of symmetrical uncertainty (SU) where it is considered the interaction between two or more variables, whereas SU can only consider the interaction between two variables. For instance, consider a table with two variables X1 and X2 and a third variable, Y (the class of the case), that results from the logical XOR operator applied to X1 and X2

X1	X2	Y
0	0	0
0	1	1
1	0	1
1	1	0

For this case

MSU(X1, X2, Y) = 0.5.

This, in contrast to the measurements obtained by SU of the variables X1 and X2 against Y,

SU(X1, Y) = 0

and

SU(X2, Y) = 0.

Usage

msu(table_variables, table_class)

Arguments

table_variables	A list of factors as categorical variables.
table_class	A factor representing the class of the case.

Value

Multivariate symmetrical uncertainty estimation for the variable set {table_variables, table_class}. The result is rounded to 7 decimal places.

See Also

symmetrical_uncertainty

Examples

# completely predictable
msu(list(factor(c(0,0,1,1))), factor(c(0,0,1,1)))
# XOR
msu(list(factor(c(0,0,1,1)), factor(c(0,1,0,1))), factor(c(0,1,1,0)))
## Not run: 
msu(c(factor(c(0,0,1,1)), factor(c(0,1,0,1))), factor(c(0,1,1,0)))
msu(list(factor(c(0,0,1,1)), factor(c(0,1,0,1))), c(0,1,1,0))

## End(Not run)

msu documentation built on May 2, 2019, 6:43 a.m.