This shows you the differences between two versions of the page.

Link to this comparison view

blogs:pub2005:genetic_process_mining_a_basic_approach_and_its_challenges [2009/05/25 12:24] (current)
Line 1: Line 1:
 +====== Genetic Process Mining: A Basic Approach and its Challenges ======
 +A.K. Alves de Medeiros, A.J.M.M.Weijters and W.M.P. van der Aalst\\
 +//Workshop on Business Process Intelligence (BPI), Nancy, 2005//\\
 +===== Abstract =====
 +One of the aims of process mining is to retrieve a process
 +model from a given event log. However, current techniques have problems
 +when mining processes that contain non-trivial constructs and/or
 +when dealing with the presence of noise in the logs. To overcome these
 +problems, we try to use genetic algorithms to mine process models. The
 +non-trivial constructs are tackled by choosing an internal representation
 +that supports them. The noise problem is naturally tackled by the genetic
 +algorithm because, per definition, these algorithms are robust to noise.
 +The definition of a good fitness measure is the most critical challenge in a
 +genetic approach. This paper presents the current status of our research
 +and the pros and cons of the fitness measure that we used so far. Experiments
 +show that the fitness measure leads to the mining of process models
 +that can reproduce all the behavior in the log, but these mined models
 +may also allow for extra behavior. In short, the current version of the genetic
 +algorithm can already be used to mine process models, but future
 +research is necessary to always ensure that the mined models do not allow
 +for extra behavior. Thus, this paper also discusses some ideas for future
 +research that could ensure that the mined models will always only reflect
 +the behavior in the log.
 +===== Links =====
 +{{publications:​Medeiros2005.pdf|Download PDF}} (281 KB)