randomForest
.## S3 method for class 'randomForest':
importance(x, type=NULL, class=NULL, scale=TRUE, ...)
randomForest
class
and type
are NULL
, A matrix of
p
rows and nclass + 2
columns (where p
is the
number of variables in the data and nclass
is the number of
classes) for classification problem, or p
rows and 2
columns for regression. In the classification case, the first
nclass
columns are the class-specific importance measures
(based on permutation of out-of-bag data). The nclass+1
st
column is the overall importance, and the last column is the overall
measure based on the decrease in node purity (or `Gini-based'
measure). If either type
or class
is given, it returns a
(named) vector of importance measure, one for each predictor variable.
randomForest
for explanation
of how the importance measures are computed.randomForest
, varImpPlot