subset-methods: Methods for Function 'subset' in Package...

In performanceEstimation: An Infra-Structure for Performance Estimation of Predictive Models

Description

The method subset when applied to objects of class ComparisonResults can be used to obtain another object of the same class with a subset of the estimation results contained in the original object.

Methods

signature(x = "ComparisonResults")

The method has as first argument the object of class ComparisonResults that you wish to subset. This method also includes 4 extra arguments that you can use to supply the subsetting criteria.

Namely, the parameter metrics allows you to indicate a vector with the subset of evaluation metrics in the orginal object. Alternatively, you can provide a regular expression to be matched against the name of the statistics measured in the experiment (see function metricNames) to specify the subset of metrics you want to select.

The parameter workflows can be used to provide a vector with the subset of workflows (approaches to the predictive tasks) that are to be used in the sub-setting. Alternatively, you can also provide a regular expression to be matched against the name of the workflows (see function workflowNames) evaluated in the experiment to specify the subset you want to select.

The parameter tasks allows you to indicate a vector with the subset of predictive tasks to be extracted. Alternatively, you can provide a regular expression to be matched against the name of the tasks used in the experiment (see function taskNames) to specify the subset you want to select.

Finally, the parameter partial allows you to control how the names of the other parameters (tasks, workflows and metrics) are matched against the original names in the ComparisonResults object. It defaults to TRUE which means that the function grep is used, whilst with FALSE the mathcin is done using the function match. For instance, if you have three metrics being estimated with names "F", "F1" and "F2", and you call subset with metrics="F", this would in effect subset all three metrics, whilst with partial=FALSE only the first of the three would be considered.

performanceEstimation documentation built on May 2, 2019, 6:01 a.m.