Fuzzy preprocessing of medical text annotations of intensive care units patients

Physicians´ and nurses´ text notes are a direct representation of experts´ views on real time observable data, and contain rich information that can improve predictive medic decision making models based on physiological data. However, such notes are not usually used in such decision making models because of the dificulties in processing their content. This paper presents a new fuzzy based text preprocessing technique that allows for a more effective use of Natural Language Processing techniques when approaching medical text reports. The technique was developed to address the reports contained in the MIMIC II database.