VSURF (version 0.6)

VSURF.interp.tune: Tuning of the interpretation step of VSURF

Description

This function allows to tune the "interpretation step" of VSURF, without rerunning all computations.

Usage

VSURF.interp.tune(res.interp, nsd = 1)

Arguments

res.interp

An object of class VSURF.interp, which is the result of the VSURF.interp function.

nsd

Number of times the standard deviation of the minimum value of err.interp is multiplied. See details below.

Value

• A list with the following components:
• varselect.interp
  A vector of indices of selected variables.
• err.interp
  A vector of the mean OOB error rates of the embedded random forests models.
• sd.min
  The standard deviation of OOB error rates associated to the random forests model attaining the minimum mean OOB error rate.
• num.varselect.interp
  The number of selected variables.
• varselect.thres
  A vector of indexes of variables selected after "thresholding step", sorted according to their mean VI, in decreasing order.

Details

In VSURF.interp function, the smallest model (and hence its corresponding variables) having a mean OOB error rate less than err.min + nsd * sd.min is selected. The function VSURF.interp.tune allows to change the value of nsd (which multiply the standard deviation of the minimum OOB error rate sd.min), without rerunning all computations. To get a larger model than default, choose a value of nsd less than 1, and to get a smaller one, choose a value larger than 1.

References

Genuer, R. and Poggi, J.M. and Tuleau-Malot, C. (2010), Variable selection using random forests, Pattern Recognition Letters 31(14), 2225-2236

See Also

VSURF, VSURF.interp

Examples

data(iris)
iris.thres <- VSURF.thres(x=iris[,1:4], y=iris[,5], ntree=100, nfor.thres=20)
iris.interp <- VSURF.interp(x=iris[,1:4], y=iris[,5], vars=iris.thres$varselect.thres,
                            nfor.interp=10)
iris.interp
iris.interp.tuned <- VSURF.interp.tune(res.interp=iris.interp, nsd=10)
iris.interp.tuned