bootEstimates {performanceEstimation}

Performance estimation using (e0 or .632) bootstrap

Description

This function obtains boostrap estimates of performance metrics for a given predictive task and method to solve it (i.e. a workflow). The function is general in the sense that the workflow function that the user provides as the solution to the task, can implement or call whatever modeling technique the user wants.

The function implements both e0 boostrap estimates as well as .632 boostrap. The selection of the type of boostrap is done through the estTask argument (check the help page of Bootstrap).

Please note that most of the times you will not call this function directly, though there is nothing wrong in doing it, but instead you will use the function performanceEstimation, that allows you to carry out performance estimation for multiple workflows on multiple tasks, using some estimation method like for instance boostrap. Still, when you simply want to have the boostrap estimate for one workflow on one task, you may use this function directly.

Usage

bootEstimates(wf,task,estTask,cluster)

Arguments

wf

an object of the class Workflow representing the modeling approach to be evaluated on a certain task.

task

an object of the class PredTask representing the prediction task to be used in the evaluation.

estTask

an object of the class EstimationTask indicating the metrics to be estimated and the boostrap settings to use.

cluster

an optional parameter that can either be TRUE or a cluster. In case of TRUE the function will run in parallel and will internally setup the parallel back-end (defaulting to using half of the cores in your local machine). You may also setup outside your parallel back-end (c.f. makeCluster) and then pass the resulting cluster object to this function using this parameter. In case no value is provided for this parameter the function will run sequentially.

Details

The idea of this function is to carry out a bootstrap experiment with the goal of obtaining reliable estimates of the predictive performance of a certain modeling approach (denoted here as a workflow) on a given predictive task. Two types of bootstrap estimates are implemented: i) e0 bootstrap and ii) .632 bootstrap. Bootstrap estimates are obtained by averaging over a set of k scores each obtained in the following way: i) draw a random sample with replacement with the same size as the original data set; ii) obtain a model with this sample; iii) test it and obtain the estimates for this run on the observations of the original data set that were not used in the sample obtained in step i). This process is repeated k times and the average scores are the bootstrap estimates. The main difference between e0 and .632 bootstrap is the fact that the latter tries to integrate the e0 estimate with the resubstitution estimate, i.e. when the model is learned and tested on the full available data sample.

Parallel execution of the estimation experiment is only recommended for minimally large data sets otherwise you may actually increase the computation time due to communication costs between the processes.

Value

The result of the function is an object of class EstimationResults.

Author(s)

Luis Torgo [e-mail]

References

Torgo, L. (2014) An Infra-Structure for Performance Estimation and Experimental Comparison of Predictive Models in R. arXiv:1412.0436 [cs.MS] http://arxiv.org/abs/1412.0436

See Also

Bootstrap, Workflow, standardWF, PredTask, EstimationTask, performanceEstimation, hldEstimates, loocvEstimates, cvEstimates, mcEstimates, EstimationResults

Examples

## Not run: 

## Estimating the MSE of a SVM variant on the 
##  swiss data, using 50 repetitions of .632 bootstrap
library(e1071)
data(swiss)

## running the estimation experiment
res <- bootEstimates(
  Workflow(wfID="svmC10G01",
           learner="svm",learner.pars=list(cost=10,gamma=0.1)
          ),
  PredTask(Infant.Mortality ~ .,swiss),
  EstimationTask("mse",method=Bootstrap(type=".632",nReps=50))
  )

## Check a summary of the results
summary(res)


## End(Not run)