Are we following the expected process?
With conformance checking, people aim to compare the behavior that
occurs in reality with the one that is captured in the process
model.
What is Conformance Checking?
Use case
There are three main use cases:
-
Compliance checking (for auditing, fraud
detection, etc.):
- Audits are performed to ascertain the validity and reliability of information about organizations and their associated processes.
- The audits help in checking whether business processes are executed within certain boundaries set by managers, governments, and other stakeholders.
-
Evaluating process discovery results/algorithms:
- Comparing discovered process models with the data used to learn the model or with unseen test data.
- Evaluating a model or an algorithm (see k-fold cross-validation).
- Conformance to specification (software, services, etc.)
Conformance checking algorithm
Token Replay
Token-replay is a heuristic technique, which uses four counters
(produced tokens, consumed tokens, missing tokens, and remaining
tokens) to compute the fitness of an
observation trace based on a given process model. Although the
approach is easy to understand and can be implemented efficiently,
since token-replay takes a local decision, it may lead to misleading
results.
Alignments
Alignments is a technique, which performs an exhaustive search to
find out the
optimal alignment between the
observed trace and the process model. Hence, it is guaranteed to
return the closest model run in comparison to the trace.
"A nice analogy that tells the difference between token-replay and alignments is searching for a particular place (e.g., a restaurant) in a city: in token-replay, you decide the direction to take just by looking at what you see. With alignments, you take your mobile phone and look at Google Maps, which will tell the optimal route (but pays the price of connecting to a GPS, download the city map, etc. …)"