control.errorest

Some parameters that control the behaviour of <code>errorest</code>.

misc

Improved predictive models by indirect classification and
bagging for classification, regression and survival problems
as well as resampling based estimators of prediction error.

Torsten Hothorn

ipred

Improved Predictors

Andrea Peters

Brian D. Ripley

Terry Therneau

Beth Atkinson

control.errorest function

<dl> <dt>k</dt>
<dd>integer, specify $k$ for $k$-fold cross-validation.</dd> <dt>nboot</dt>
<dd>integer, number of bootstrap replications.</dd> <dt>strat</dt>
<dd>logical, if <code>TRUE</code>, cross-validation is performed 
 using stratified sampling (for classification problems).</dd> <dt>random</dt>
<dd>logical, if <code>TRUE</code>, cross-validation is performed using
 a random ordering of the data.</dd> <dt>predictions</dt>
<dd>logical, indicates whether the prediction
 for each observation should be returned or not
 (classification and regression only). For a bootstrap
based estimator a matrix of size 'number of observations' times nboot is
returned with predicted values of the ith out-of-bootstrap sample in column
i and 'NA's for those observations not included in the ith out-of-bootstrap 
sample.</dd> <dt>getmodels</dt>
<dd>logical, indicates a list of all models should be
 returned. For cross-validation only.</dd> <dt>list.tindx</dt>
<dd>list of numeric vectors, indicating which
 observations are included in each bootstrap or cross-validation sample, respectively.</dd></dl>

Arguments

Control Error Rate Estimators — control.errorest

<dl>

 <dt>k</dt>
<dd>integer, specify $k$ for $k$-fold cross-validation.</dd>

 <dt>nboot</dt>
<dd>integer, number of bootstrap replications.</dd>

 <dt>strat</dt>
<dd>logical, if <code>TRUE</code>, cross-validation is performed 
 using stratified sampling (for classification problems).</dd>

 <dt>random</dt>
<dd>logical, if <code>TRUE</code>, cross-validation is performed using
 a random ordering of the data.</dd>

 <dt>predictions</dt>
<dd>logical, indicates whether the prediction
 for each observation should be returned or not
 (classification and regression only). For a bootstrap
based estimator a matrix of size 'number of observations' times nboot is
returned with predicted values of the ith out-of-bootstrap sample in column
i and 'NA's for those observations not included in the ith out-of-bootstrap 
sample.</dd>

 <dt>getmodels</dt>
<dd>logical, indicates a list of all models should be
 returned. For cross-validation only.</dd>

 <dt>list.tindx</dt>
<dd>list of numeric vectors, indicating which
 observations are included in each bootstrap or cross-validation sample, respectively.</dd>

</dl>

control.errorest: Control Error Rate Estimators

Description

Usage

Value

Arguments