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

Link to this comparison view

blogs:pub2008:process_mining_a_two-step_approach_to_balance_between_underfitting_and_overfitting [2009/05/25 13:43] (current)
Line 1: Line 1:
 +====== Process Mining: A Two-Step Approach to Balance Between Underfitting and Overfitting ======
 +W.M.P. van der Aalst, V. Rubin, H.M.W. Verbeek, B.F. van Dongen, E. Kindler, and C.W. Günther\\  ​
 +//Appeared as BPM Center Report BPM-08-01, BPMcenter.org,​ 2008//\\
 +===== Abstract =====
 +Process mining includes the automated discovery of processes from event logs. Based on observed events (e.g., activities being
 +executed or messages being exchanged) a process model is constructed.
 +One of the essential problems in process mining is that one cannot assume
 +to have seen all possible behavior. At best one has seen a representative
 +subset. Therefore, classical synthesis techniques are not suitable as they
 +aim at finding a model that is able to exactly reproduce the log. Existing
 +process mining techniques try to avoid such "​overfitting"​ by generalizing
 +the model to allow for more behavior. This generalization is often driven
 +by the representation language and very crude assumptions about completeness. As a result parts of the model are "​overfitting"​ (allow only
 +what has actually been observed) while other parts may be "​underfitting"​ (allow for much more behavior without strong support for it). None
 +of the existing techniques enables the user to control the balance between
 +"​overfitting"​ and "​underfitting"​. To address this, we propose a two-step
 +approach. First, using a configurable approach, a transition system is
 +constructed. Then, using the "​theory of regions",​ the model is synthesized. The approach has been implemented in the context of ProM and
 +overcomes many of the limitations of traditional approaches.
 +===== Links =====
 +  * {{publications:​bpm_08_01.pdf|Preprint of published paper}}
 +  * [[http://​www.bpmcenter.org/​|BPM Center]]