When is resampling beneficial for feature selection with imbalanced wide data?

In this repository are stored the feature selection algorithms used in article When is resampling beneficial for feature selection with imbalanced wide data? , the stored algorithms are:

• T-test
• ANOVA
• Chi squared
• Information gain
• Random forest importance
• Relief
• SVM-RFE

How to use

Run requeriments.R which will install the necessary libraries.

Rscript requeriments.R

Once the libraries are installed, inside out code we must import the functions before using them.

source("featureSelectors.R")

These functions receive as input a list with element "d" as the dataset and "tag" as its tags, function partitionDataTag can be used to format any dataframe in that way, taking into account that the tag is placed in the last column.

df = iris[1:100,]
df = partitionDataTag(df)

Finally we can launch any of the following algorithms:

fSelection_t_test(data)
fSelection_filter_anova(data)
fSelection_filter_chi.squared(data)
fSelection_filter_information.gain(data)
fSelection_random_forest_importance(data)
fSelection_filter_relief(data)
fSelection_svm_rfe(data)

The complete example can be seen on file example.R

Cite this article

@article{ramos2022resampling,
  title={When is resampling beneficial for feature selection with imbalanced wide data?},
  author={Ramos-P{\'e}rez, Ismael and Arnaiz-Gonz{\'a}lez, {\'A}lvar and Rodr{\'\i}guez, Juan J and Garc{\'\i}a-Osorio, C{\'e}sar},
  journal={Expert Systems with Applications},
  volume={188},
  pages={116015},
  year={2022},
  publisher={Elsevier}
}