NoiseFiltersR (version 0.1.0)

hybridRepairFilter: Hybrid Repair-Remove Filter

Description

Ensemble-based filter for removing or repairing label noise from a dataset as a preprocessing step of classification. For more information, see 'Details' and 'References' sections.

Usage

"hybridRepairFilter"(formula, data, ...)
"hybridRepairFilter"(x, consensus = FALSE, noiseAction = "remove", classColumn = ncol(x), ...)

Arguments

formula

A formula describing the classification variable and the attributes to be used.

data, x

Data frame containing the tranining dataset to be processed.

...

Optional parameters to be passed to other methods.

consensus

If set to TRUE, consensus voting scheme is applied to identify noisy instances. Otherwise (default), majority approach is used.

noiseAction

Character which can be set to "remove", "repair" or "hybrid". The filter accordingly decides what to do with the identified noise (see Details).

classColumn

Positive integer indicating the column which contains the (factor of) classes. By default, the last column is considered.

Value

An object of class filter, which is a list with seven components:

• cleanData is a data frame containing the filtered dataset.
• remIdx is a vector of integers indicating the indexes for removed instances (i.e. their row number with respect to the original data frame).
• repIdx is a vector of integers indicating the indexes for repaired/relabelled instances (i.e. their row number with respect to the original data frame).
• repLab is a factor containing the new labels for repaired instances.
• parameters is a list containing the argument values.
• call contains the original call to the filter.
• extraInf is a character that includes additional interesting information not covered by previous items.

Details

As presented in (Miranda et al., 2009), hybridRepairFilter builds on the dataset an ensemble of four classifiers: SVM, Neural Network, CART, KNN (combining k=1,3,5). According to their predictions and majority or consensus voting schemes, a subset of instances are labeled as noise. These are removed if noiseAction equals "remove", their class is changed into the most voted among the ensemble if noiseAction equals "repair", and when the latter is set to "hybrid", the vote of KNN decides whether remove or repair.

All this procedure is repeated while the accuracy (over the original dataset) of the ensemble trained with the processed dataset increases.

References

Miranda A. L., Garcia L. P. F., Carvalho A. C., Lorena A. C. (2009): Use of classification algorithms in noise detection and elimination. In Hybrid Artificial Intelligence Systems (pp. 417-424). Springer Berlin Heidelberg.

Examples

# Next example is not run in order to save time
## Not run: 
# data(iris)
# out <- hybridRepairFilter(iris, noiseAction = "hybrid")
# summary(out, explicit = TRUE)
# ## End(Not run)