Real-Time Fuzzy Linguistic Analysis of Anomalies from Medical Monitoring Devices on Data Streams

Analysis of data streams generated from devices collecting data from patients, which are monitored within both clinical and home environments, provide useful information for decision making processes. Nevertheless, medical personnel are still required to review and process the data and therefore spend a lot of time and effort to detect situations of concern such as exacerbations with conditions or the occurrence of anomalies in the measurements. In this paper, we propose a methodology for the real-time linguistic analysis of data streams generated from medical monitoring devices based on a rule-based inference engine exploiting a fuzzy linguistic approach. A case study based on health data provided by the Physiological Data Modeling Contest is used to illustrate the proposed methodology and to demonstrate the flexibility to interpret, in a linguistic manner, data streams and the detection of risk situations of interest based on linguistic expressions.