bsnullinteract generates bootstrapped null interaction models,
which can be used to derive a reference distribution of the test statistic
calculated with interact.
A list of length nsamp with null interaction models, to be
used as input for interact.
Arguments
object
object of class pre.
nsamp
numeric. Number of bootstrapped null interaction models to be
derived.
parallel
logical. Should parallel foreach be used to generate initial
ensemble? Must register parallel beforehand, such as doMC or others.
penalty.par.val
character or numeric. Value of the penalty parameter
\(\lambda\) to be employed for selecting the final ensemble. The default
"lambda.min" employs the \(\lambda\) value within 1 standard
error of the minimum cross-validated error. Alternatively,
"lambda.min" may be specified, to employ the \(\lambda\) value
with minimum cross-validated error, or a numeric value \(>0\) may be
specified, with higher values yielding a sparser ensemble. To evaluate the
trade-off between accuracy and sparsity of the final ensemble, inspect
pre_object$glmnet.fit and plot(pre_object$glmnet.fit).
verbose
logical. should progress be printed to the command line?
...
Further arguments to be passed to predict.pre.
Details
Note that computation of bootstrapped null interaction models is
computationally intensive. The default number of samples is set to 10,
but for reliable results argument nsamp should be set to a higher
value (e.g., \(\ge 100\)).
See also section 8.3 of Friedman & Popescu (2008).
References
Fokkema, M. (2020). Fitting prediction rule ensembles with R
package pre. Journal of Statistical Software, 92(12), 1-30.
tools:::Rd_expr_doi("10.18637/jss.v092.i12")
Friedman, J. H., & Popescu, B. E. (2008). Predictive learning
via rule ensembles. The Annals of Applied Statistics, 2(3), 916-954,
tools:::Rd_expr_doi("10.1214/07-AOAS148").