This repository contains the code used for the "Efficient Subgroup Discovery through Auto-Encoding" paper. It includes the following files:
- 'ExperimentCode.ipynb' contains code used for performing the experiment.
- 'dataImporter.py' provides a data import function which is used to load the datasets used in the paper. These datasets can be found in the 'data' folder.
- 'autoEncoder.py' provides a function to perform dimension reduction using an auto-encoder.
- 'beamSearch.ipynb', 'stdPysubgroup' and 'adjPysubgroup.py' contain the code used to run subgroup discovery algorithms. Note that 'stdPysubgroup' was not used for the experiment described in the paper.
- 'qualityMeasures.py' provides functions to evaluate the subgroups generated by the subgroup discovery algorithms.
- 'utilityFunctions.py' provides a function to perform normalization.