La r-confiance pour l'identification de trajectoires de patients
In EGC 2016, vol. RNTI-E-30, pp.535-536
Sequential patterns mining consist in identifying frequent sequences of ordered events. To solve the problem of the large number of patterns obtained, we extend the interest measure called confidence, conventionally used to select association rules to sequential patterns. We focused on a case study: myocardial infarction (MI), in order to predict the trajectory of patients with MI between 2009 and 2013. The results were submitted to an expert for discussion and validation.