Gini index for a given split of a categorical variable

Source: R/Gini.R

Calculates the Gini index of a categorical variable for a given split (threshold) value of a numeric variable.

Gini(x, group, thres)

Arguments

x

A numeric vector
group

Categorical variable
thres

Threshold value to use for the split

Value

A data frame with Gini index before and after the split, as well as the gain and gain percent

References

Kuhn, M. & Johnson, K. (2013). Applied Predictive Modeling. Springer.

Witten, I. H., Frank, E. & Hall, M. A. (2011). Data Mining: Practical Machine Learning Tools and Techniques. Elsevier.

Examples

Gini(mtcars$mpg, mtcars$cyl, 15)

#>  gini.init gini.split gini.gain gini.gain.perc
#>      0.643      0.553      0.09             14
Gini(mtcars$mpg, mtcars$cyl, 21)

#>  gini.init gini.split gini.gain gini.gain.perc
#>      0.643      0.342     0.301           46.8
Gini(mtcars$mpg, mtcars$cyl, 25)

#>  gini.init gini.split gini.gain gini.gain.perc
#>      0.643      0.488     0.155           24.1