summary.filter: Summary method for class filter

In NoiseFiltersR: Label Noise Filters for Data Preprocessing in Classification

Description

This methods allows for appropriately displaying the most important information about a filtered dataset, contained in the S3 class filter.

Usage

## S3 method for class 'filter'
summary(object, ..., explicit = FALSE)

Arguments

object	Object of class filter.
...	Additional arguments affecting the summary produced.
explicit	If set to TRUE, the indexes for removed and repaired instances (as well as new labels for the latters) are displayed. It defaults to FALSE.

Details

The information offered is the following:

• Names of the dataset and the filter.

• Original call to the filter.

• Specific parameters used for the filter.

• Results: number of removed and repaired instances (absolute number and percentage).

• Additional information (if available, it depends on the filter).

• Optionally, if explicit=TRUE, the indexes for removed and repaired instances, as well as the new labels.

Examples

# Next example is not run in order to save time
## Not run: 
# Example of filter with additional information available.
data(iris)
out <- edgeBoostFilter(Species~., data = iris)
class(out)
summary(out)
summary(out, explicit = TRUE)

## End(Not run)

Example output

OpenJDK 64-Bit Server VM warning: Can't detect primordial thread stack location - find_vma failed
[1] "filter"
Filter edgeBoostFilter applied to dataset iris 

Call:
edgeBoostFilter(formula = Species ~ ., data = iris)

Parameters:
m: 15
percent: 0.05
threshold: 0

Results:
Number of removed instances: 9 (6 %)
Number of repaired instances: 0 (0 %)

Additional information:
Highest edge value kept: 0.0669425282293152 
Filter edgeBoostFilter applied to dataset iris 

Call:
edgeBoostFilter(formula = Species ~ ., data = iris)

Parameters:
m: 15
percent: 0.05
threshold: 0

Results:
Number of removed instances: 9 (6 %)
Number of repaired instances: 0 (0 %)

Additional information:
Highest edge value kept: 0.0669425282293152 

Explicit indexes for removed instances:
58 64 78 84 107 120 130 134 139 

NoiseFiltersR documentation built on May 2, 2019, 2:03 a.m.