Similarity-based Personalized Risk Calculation

In patient monitoring systems, personalized risk calculation is of vital importance. Instead of the generalized medical recommendations defined in literature, patient-specific limits should be applied, taking into account the specific medical history of the given user. To follow up and analyse the patient's condition, the measured values have to be compared to the specified recommendations. During this comparison it is insufficient to examine the instantaneous value, it is necessary to take into account the measurement statistics for longer periods. In this paper statistics-based fuzzy sets are evaluated using similarity measures. The basic idea of the proposed method is to create a fuzzy set representing the current state of the patient based on the histogram of the measured values. The authors propose a Simplified Centre of Gravity method-based similarity measure to compare the current state to the personal medical recommendations. The aim is to assess patient state progress, based on which the improvement or deterioration of the overall health condition can be monitored.