This is ongoing research. The extended abstract as well as the poster presentation show our preliminary results. Both were presented at [IC2S2 2019][1] in July 2019 in Amsterdam. The R scripts allowing a closer look at the modeling to achieve compression will be made available as soon as development has reached the next stage. The project is based on work by developmental scientists at [Leiden University] [2], Netherlands in collaboration with the University of Evora, Portugal. We use information theoretic data mining to explore subgroups in the data. The aim is to ultimately use raw RFID data and integrate noise reduction into the actual subgroup mining method. This will remove explicit parameter setting at the pre-processing of the data. Note however that the model class needs to be determined prior to mining subgroups. More information on the work of the Explanatory Data Analysis group at the Leiden Institute of Advanced Computer Science can be found [here][3].   [1]: https://2019.ic2s2.org [2]: https://www.focusonemotions.nl/projects/play [3]: http://eda.liacs.nl