C5imp: Variable Importance Measures for C5.0 Models

Description

This function calculates the variable importance (aka attribute usage) for C5.0 models.

Usage

C5imp(object, metric = "usage", pct = TRUE, ...)

Arguments

object

an object of class C5.0

metric

either 'usage' or 'splits' (see Details below)

pct

a logical: should the importance values be converted to be between 0 and 100?

...

other options (not currently used)

Value

a data frame with a column Overall with the predictor usage values. The row names indicate the predictor.

Details

By default, C5.0 measures predictor importance by determining the percentage of training set samples that fall into all the terminal nodes after the split (this is used when metric = "usage"). For example, the predictor in the first split automatically has an importance measurement of 100 percent. Other predictors may be used frequently in splits, but if the terminal nodes cover only a handful of training set samples, the importance scores may be close to zero. The same strategy is applied to rule-based models as well as the corresponding boosted versions of the model.

There is a difference in the attribute usage numbers between this output and the nominal command line output. Although the calculations are almost exactly the same (we do not add 1/2 to everything), the C code does not display that an attribute was used if the percentage of training samples covered by the corresponding splits is very low. Here, the threshold was lowered and the fractional usage is shown.

When metric = "splits", the percentage of splits associated with each predictor is calculated.

References

Quinlan R (1993). C4.5: Programs for Machine Learning. Morgan Kaufmann Publishers, http://www.rulequest.com/see5-unix.html

Examples

Run this code

data(churn)

treeModel <- C5.0(x = churnTrain[, -20], y = churnTrain$churn)
C5imp(treeModel)
C5imp(treeModel, metric = "splits")

Run the code above in your browser using DataLab