Calculation of variable importance for regression and classification models
A generic method for calculating variable importance for objects produced by
train and method specific methods
## S3 method for class 'train': varImp(object, useModel = TRUE, nonpara = TRUE, scale = TRUE, ...)
## S3 method for class 'earth': varImp(object, value = "gcv", ...)
## S3 method for class 'fda': varImp(object, value = "gcv", ...)
## S3 method for class 'rpart': varImp(object, surrogates = FALSE, competes = TRUE, ...)
## S3 method for class 'randomForest': varImp(object, ...)
## S3 method for class 'gbm': varImp(object, numTrees, ...)
## S3 method for class 'classbagg': varImp(object, ...)
## S3 method for class 'regbagg': varImp(object, ...)
## S3 method for class 'pamrtrained': varImp(object, threshold, data, ...)
## S3 method for class 'lm': varImp(object, ...)
## S3 method for class 'mvr': varImp(object, estimate = NULL, ...)
## S3 method for class 'bagEarth': varImp(object, ...)
## S3 method for class 'bagFDA': varImp(object, ...)
## S3 method for class 'RandomForest': varImp(object, ...)
## S3 method for class 'rfe': varImp(object, drop = FALSE, ...)
## S3 method for class 'dsa': varImp(object, cuts = NULL, ...)
## S3 method for class 'multinom': varImp(object, ...)
## S3 method for class 'cubist': varImp(object, weights = c(0.5, 0.5), ...)
## S3 method for class 'JRip': varImp(object, ...)
## S3 method for class 'PART': varImp(object, ...)
## S3 method for class 'C5.0': varImp(object, ...)
## S3 method for class 'nnet': varImp(object, ...)
## S3 method for class 'glmnet': varImp(object, lambda = NULL, ...)
## S3 method for class 'plsda': varImp(object, ...)
- an object corresponding to a fitted model
- use a model based technique for measuring variable importance? This is only used for some models (lm, pls, rf, rpart, gbm, pam and mars)
- should nonparametric methods be used to assess the relationship
between the features and response (only used with
useModel = FALSEand only passed to
- should the importance values be scaled to 0 and 100?
- parameters to pass to the specific
- the number of iterations (trees) to use in a boosted tree model
- the shrinkage threshold (
- the training set predictors (
- the statistic that will be used to calculate importance:
- should surrogate splits contribute to the importance calculation?
- should competing splits contribute to the importance calculation?
- which estimate of performance should be used? See
- a logical: should variables not included in the final set be calculated?
- the number of rule sets to use in the model (for
- a numeric vector of length two that weighs the usage of variables in the rule conditions and the usage in the linear models (see details below).
- a single value of the penalty parameter
For models that do not have corresponding
varImp methods, see
Linear Models: the absolute value of the t--statistic for each model parameter is used.
varImp.RandomForest are wrappers around the importance functions from the
rpart.control. This method does not currently provide
class--specific measures of importance when the response is a factor.
Bagged Trees: The same methodology as a single tree is applied to
all bootstrapped trees and the total importance is returned
varImp.gbm is a wrapper around the function from that package (see the
varImp function tracks the changes in
model statistics, such as the GCV, for each predictor and
accumulates the reduction in the statistic when each
predictor's feature is added to the model. This total reduction
is used as the variable importance measure. If a predictor was
never used in any of the MARS basis functions in the final model
(after pruning), it has an importance
value of zero. Prior to June 2008, the package used an internal function
for these calculations. Currently, the
varImp is a wrapper to
evimp function in the
earth package. There are three statistics that can be used to
estimate variable importance in MARS models. Using
varImp(object, value = "gcv") tracks the reduction in the
generalized cross-validation statistic as terms are added.
However, there are some cases when terms are retained
in the model that result in an increase in GCV. Negative variable
importance values for MARS are set to zero.
varImp(object, value = "rss") monitors the change in the
residual sums of squares (RSS) as terms are added, which will
never be negative.
Also, the option
varImp(object, value =" nsubsets"), which
counts the number of subsets where the variable is used (in the final,
Nearest shrunken centroids: The difference between the class centroids and the overall centroid is used to measure the variable influence (see
pamr.predict). The larger the difference between the class centroid and the overall center of the data, the larger the separation between the classes. The training set predictions must be supplied when an object of class
pamrtrained is given to
Cubist: The Cubist output contains variable usage
statistics. It gives the percentage of times where each variable was
used in a condition and/or a linear model. Note that this output
will probably be inconsistent with the rules shown in the output
summary.cubist. At each split of the
tree, Cubist saves a linear model (after feature selection) that is
allowed to have terms for each variable used in the current split or
any split above it. Quinlan (1992) discusses a smoothing algorithm
where each model prediction is a linear combination of the parent
and child model along the tree. As such, the final prediction is a
function of all the linear models from the initial node to the
terminal node. The percentages shown in the Cubist output reflects
all the models involved in prediction (as opposed to the terminal
models shown in the output). The variable importance used here is a
linear combination of the usage in the rule conditions and the
PART and JRip: For these rule-based models, the
importance for a predictor is simply the number of rules that
involve the predictor.
C5.0: C5.0 measures predictor importance by determining the
percentage of training set samples that fall into all the terminal
nodes after the split. For example, the predictor in the first split
automatically has an importance measurement of 100 percent since all
samples are affected by this split. 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 and boosted
versions of the model. The underlying function can also return the
number of times each predictor was involved in a split by using the
metric = "usage".
Neural Networks: The method used here is based on Gevrey et al (2003), which uses combinations of the absolute values of the weights. For classification models, the class-specific importances will be the same. Recursive Feature Elimination: Variable importance is computed using the ranking method used for feature selection. For the final subset size, the importances for the models across all resamples are averaged to compute an overall value.
Feature Selection via Univariate Filters, the percentage of resamples that a predictor was selected is determined. In other words, an importance of 0.50 means that the predictor survived the filter in half of the resamples.
- A data frame with class
varImp.trainor a matrix for other models.
Gevrey, M., Dimopoulos, I., & Lek, S. (2003). Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecological Modelling, 160(3), 249-264.
Quinlan, J. (1992). Learning with continuous classes. Proceedings of the 5th Australian Joint Conference On Artificial Intelligence, 343-348.