Evaluation Goals for Online Process Mining : a Concept Drift Perspective
Abstract : Online process mining refers to a class of techniques for analyzing in real-time event streams generated by the execution of business processes. These techniques are crucial in the reactive monitoring of business processes, timely resource allocation and detection/prevention of dysfunctional behavior. This paper fills the gap by identifying a set of evaluation goals for online process mining and examining their fulfillment in the state of the art. Many interesting advances have been made by the research community in recent years , but there is no consensus on the exact set of properties these techniques have to achieve.We discuss parameters and techniques regulating the balance between conflicting goals and outline research needed for their improvement.Concept drift detection is crucial in this sense but, as demonstrated by our experiments, it is only partially supported by current solutions
? we have presented the first online mechanism for detecting and managing concept drift, combined with the process mining approach in based on abstract interpretation and Petri net models.
? Our experiments on process mining bench mark data twisted to incorporate drift show that our method detects abrupt changes quickly and accurately.
? We have also described how to apply the mechanism for a richer set of tasks: characterizing and locating change , unraveling the process change, and revising the mined models
? we use an adaptive window technique which has been proved to be very effective for similar purposes
? The problem of bounding the memory usage is not addressed (7G1), moreover, the time of accumulation implies a delay in the response (7G2)
? In order to address the problem of computing the conformance of incomplete cases, in an approach for assessing the optimal alignment of sub-traces is proposed.
? He authors are aware of this problem and propose either Random Sampling with Reservoir or Decay-based data structures similarly to the solutions provided respectively in for process discovery.
? This way it is possible to manage the trade-off between memory consumption (3G1) and accuracy(–G4). It is clear that CDD in non-stationary environments is a precondition to not lose accuracy.
? we present the first online mechanism for detecting and managing concept drift, which is based on abstract interpretation and sequential sampling, together with recent learning techniques on data streams.
? we concentrate on the control-flow part, i.e., the causal relations between the events of a process.
? This paper presents an online technique to detect concept drift by sequential monitoring of the logs of a system. It is a multi-stage technique that uses the theory of abstract interpretation
? The amount of memory that can be used during data analysis is much smaller than the entire series
? The events accumulated are used to generate an abstract interpretation of their behavior using a convex polyhedron that offers an upper approximation.
? When new cases are acquired , online analysis matches them with the polyhedron to estimate their divergence to the previously observed behavior, supporting the online assessment of concept drift
? To improve the accuracy of change-point detection, statistical tests were introduced to set the optimal size of the window of analysis but disregarding then memory management.
|