Subgroup Discovery with Evolutionary Fuzzy Systems

Implementation of evolutionary fuzzy systems for the data mining task called "subgroup discovery". In particular, the algorithms presented in this package are: M. J. del Jesus, P. Gonzalez, F. Herrera, M. Mesonero (2007) M. J. del Jesus, P. Gonzalez, F. Herrera (2007) C. J. Carmona, P. Gonzalez, M. J. del Jesus, F. Herrera (2010) C. J. Carmona, V. Ruiz-Rodado, M. J. del Jesus, A. Weber, M. Grootveld, P. González, D. Elizondo (2015) It also provide a Shiny App to ease the analysis. The algorithms work with data sets provided in KEEL, ARFF and CSV format and also with data.frame objects.




  • Change the name of the read.keel() function to read.dataset().
  • Added direct support for CSV files on read.dataset().
  • Rename keel class to SDEFSR_Dataset class.
  • Rename keelFromDataFrame() to SDEFSR_DatasetFromDataFrame().
  • Added a new object SDEFSR_Rules which:
    • Contains all rules generated by an SD algorithm an its associated quality measures.
    • The "[]" operator supports filter rules by quality measure
    • Added S3 method "sort" to sort rules by a given quality measure
  • Added new method plotRules() which shows a TPR vs FPR plot by using ggplot2 package.
  • Fixed error when reading ARFF files with quoted '' variables.
  • Changes of the UI:
    • Added a new visualization method: Variable vs variable
    • Added functionality to filter numerical variables
    • Added functionality to filter instences by the given numeric or categorical filters
    • On quality measure tab, the functionality to show the plot of plotRules() is displayed
    • Changed the server logic to work with the new SDEFSR_Rules object to improve performance on large results.


  • Added a new subgroup discovery algorithm: FuGePSD.
  • New dataset files support:
    • ARFF files support from read.keel()
    • Conversion from data.frame to keel with keelFromDataFrame() function.
  • Minor optimizations when reading a file.
  • Some algorithm optimizations, changed a bit the way of calling the functions Be aware with this!
  • Fixed problem of the algorithms that returns four files instead of three. Now all the algorithms (except FuGePSD) returns three files.
  • Created S3 method to visualize keel objects with print() and summarize with summary()
  • Changes of the UI:
    • Algorithm selection an parameters now are on the tab panel of the right side of the UI after "Exploratory analysis". We think with this change the user has a better workflow and a better visualitation and organization of the results.
    • Also, the new functionality is available on the GUI, except the conversion from data.frame.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.7.21 by Angel M. Garcia, 25 days ago

Browse source code at

Authors: Angel M. Garcia [aut, cre] , Pedro Gonzalez [aut, cph] , Cristobal J. Carmona [aut, cph] , Francisco Charte [ctb] , Maria J. del Jesus [aut, cph]

Documentation:   PDF Manual  

LGPL (>= 3) license

Imports grDevices, methods, parallel, stats, utils

Suggests ggplot2, knitr, shiny

See at CRAN