SubDisc is a Data Mining tool for discovering local patterns in data. SubDisc features a generic Subgroup Discovery algorithm that can be configured in many ways, in order to implement various forms of local pattern discovery. The tool can deal with a range of data types, both for the input attributes as well as the target attributes, including nominal, numeric and binary.

A unique feature of SubDisc is its ability to deal with a range of Subgroup Discovery settings, determined by the type and number of target attributes. Where regular SD algorithms only consider a single target attribute, nominal or sometimes numeric, Cortana is able to deal with targets consisting of multiple attributes, in a setting called Exceptional Model Mining.

SubDisc was previously developed under the name Cortana.

• Generic parameterized Subgroup Discovery algorithm.
• Multiple data types supported.
• Implemented in Java, so works on all major platforms, including Windows, Linux and Mac OS.
• Works on propositional (tabular) data from flat files, .TXT or .ARFF.
• Includes Exceptional Model Mining settings.
• Statistical validation of mining results.
• Graphical presentation of results, such as ROC curves, scatter plots, and exceptional models.
• Additional bioinformatics module for literature-based gene set enrichment (see bioinformatics below).
• Free binary version and open-source access.
• Wrapper available for R (https://github.com/SubDisc/rSubDisc) and Python (soon)

The code is compatible with Java 15.

1. Either download the last released version jar file (https://github.com/SubDisc/SubDisc/releases/) or build it yourself (below).
2. Double-click on the .jar file or use java cli (ex.: java -jar target/subdisc-gui.jar).

The interface should appear, and you are ready to open a data file and discover subgroups!

1. Clone the repository: git clone https://github.com/SubDisc/SubDisc.git
2. Use maven to assemble the .jar file: mvn package
3. The .jar file is created in ./target and named something like subdisc-gui-2.1094.jar.

Technical details concerning the algorithms behind Cortana can be found in various scientific publications:

• Subgroup Discovery in Ranked Data, with an Application to Gene Set Enrichment
• Exceptional Model Mining
• Exploiting False Discoveries - Statistical Validation of Patterns and Quality Measures in Subgroup Discovery
• Diverse Subgroup Set Discovery
• Flexible Enrichment with Cortana (Software Demo)
• Efficient Algorithms for Finding Richer Subgroup Descriptions in Numeric and Nominal Data
• Non-Redundant Subgroup Discovery in Large and Complex Data
• Subgroup Discovery meets Bayesian networks: an Exceptional Model Mining approach
• Discovering Local Subgroups, with an Application to Fraud Detection
• A Bayesian Scoring Technique for Mining Predictive and Non-Spurious Rules

The following people have contributed in various ways to the development of SubDisc/Cortana:

• Arno Knobbe
• Marvin Meeng
• Wouter Duivesteijn
• atthijs van Leeuwen
• Rob Konijn
• Michael Mampaey
• Siegfried Nijssen
• Iyad Batal
• Jeremie Gobeil