Temporal abstractions to enrich Activity-Based Process Mining corpus with clinical time series

The design of clinical protocols for improving the quality of care in an efficient way is one of the challenges for the deployment of Evidence Based Medicine. The design of those protocols is a difficult task that require the consensus of care process experts. The use of Pattern Recognition approaches, like Process Mining, allows the automatic inference of processes that can help experts for formalizing these clinical protocols based on the actually deployed care process. However, the step rules among the different stages of the care protocols are based on high level descriptions of numerical clinical data gathered from the patient that can not be processed directly by Process Mining approaches. In this paper, a combination of Interactive Pattern Recognition with Temporal Abstraction technologies that allows processing of clinical data to allow the enrichment of Activity Based Process Mining corpus is presented.