Process Model Plugins in Pro M

Hi,

A brief introduction to a number of typical miners.

Alpha Miner: The first miner, with formal guarantees using some assumptions. Can however lead to unsound process models if the assumptions are not met. I would not use this miner on real life logs.

Heuristics Miner: A simple miner using far less assumptions. Will work in general, but results in a heuristics net. Several conversions from a heuristics net to a Petri net have been implemented, but not all implementations are okay.

ILP Miner: A miner that always results in a Petri net that fits the event log perfectly, meaning that all traces can be replayed successfully on the net. Very sensitive to noise, as all traces need to be replayable in the net. I would not use this miner on real-life logs.

Inductive Miner: A miner that always discovers a sound Petri net, and that works in general. If you have no other clues, use this miner. It can handle noise. The miner may result in subnets that correspond to a 'flower' model. If the miner does not know, it will create a subnet in which anything can be replayed successfully.

Kind regards,

Eric.