Process Mining

We are are developing a two-step approach in the area of process mining using synthesis algorithms for p/t-nets:

  • First step (preprocessing): Construct an labelled prime event structure (LPES) with cutoff-list from an event log through detecting loops, causality, concurrency and noise. Since LPES with cutoff-list are a more general model than p/t-nets, causal structures can be detected with less loss of information than with approaches using directly p/t-net based models. For the handling of noise we employ statistical methods.
  • Second step (synthesis): Synthesize a Petri net from the result of the first step.


We developed command line tool in JAVA for the synthesis of bounded p/t-nets from LPES with cutoff-list using wrong continuations. We are currently reimplementing the tool in order to improve its performance. In the meanwhile, no download is available.   




We are currently working on using hypothesis tests in order to filter infrequent behavior from an event log. We developed the command line tool h0logfilter for producing a causal footprint without infrequent behavior from an event log. The footprint can be used to construct directly a process model from it or to construct an LPES with cutoff-list reflecting the causal structures detected in the log.