IPDD Adaptive version 1.5

This release updates the Adaptive Approach for Control-flow drifts in the Interactive Process Drift Detection (IPDD) Framework, named IPDD Adaptive. IPDD is a tool for process drift detection and analysis.

Version 1.5 includes two approaches for control-flow drift detection: Trace by Trace and Windowing.

This release uses the Pm4Py library in version 2.7.5 and a customized version of the scikit-multiflow.

We have included the synthetic datasets used in the manufacturing experiments.

Information about new version of pm4py

We have updated the pm4py from version 2.2.20.1 to version 2.7.5. After this update the pm4py inductive miner return more precise models, which affects the precision metric value used by IPDD for detecting the concept drifts. Because of this difference, when executing the updated version of IPDD will result on different detections for some scenarios when comparing to the reported results at my thesis: https://www.ppgia.pucpr.br/pt/arquivos/doutorado/teses/2022/Tese_Denise_Maria_Vecino_Sato.pdf.

"Since the release of pm4py 2.3.0, the inductive miner has been refactored. In particular, a change of interest is the introduction of the "strict sequence cut" in place of the traditional "sequence cut". This type of sequence cut provides generally more precise models, but results in models that are different." - information provided by Alessandro Berti from pm4py team.