llama (version 0.10.1)

imputeCensored: Impute censored values

Description

Impute the performance values that are censored, i.e. for which the respective algorithm was not successful.

Usage

imputeCensored(data = NULL, estimator = makeLearner("regr.lm"),
    epsilon = 0.1, maxit = 1000)

Arguments

data

the data to check for censored values to impute. The structure returned by input.

estimator

the mlr regressor to use to impute the censored values.

epsilon

the convergence criterion. Default 0.1.

maxit

the maximum number of iterations. Default 1000.

Value

The data structure with imputed censored values. The original data is saved in the original_data member.

Details

The function checks for each algorithm if there are censored values by checking for which problem instances the algorithm was not successful. It trains a model to predict the performance value for those instances using the given estimator based on the performance values of the instances where the algorithm was successful and the problem features. It then uses the results of this initial prediction to train a new model on the entire data and predict the performance values for those problems where the algorithm was successful again. This process is repeated until the maximum difference between predictions in two successive iterations is less than epsilon or more than maxit iterations have been performed.

It is up to the user to check whether the imputed values make sense. In particular, for solver runtime data and timeouts one would expect that the imputed values are above the timeout threshold, indicating at what time the algorithms that have timed out would have solved the problem. No effort is made to enforce such application-specific constraints.

References

Josef Schmee and Gerald J. Hahn (1979) A Simple Method for Regression Analysis with Censored Data. Technometrics 21, no. 4, 417-432.

Examples

# NOT RUN {
if(Sys.getenv("RUN_EXPENSIVE") == "true") {
data(satsolvers)
imputed = imputeCensored(satsolvers)
}
# }