PREDICTING THE EMERGENCE AND RECURRENCE OF BRAIN STROKE BASED ON HYBRID FUZZY MODELS

UDC 616-005.4

N.A. Korenevskiy, A.V. Bykov, E.V. Tsymbal, V.V. Aksenov, D.S. Rodionov


The work is devoted to the actual problem of improving the quality and efficiency of predicting the occurrence and recurrence of brain stroke by using the methodology of synthesis of hybrid fuzzy mathematical models developed at the Department of biomedical engineering of Southwestern state University.In the course of the conducted research the space of informative signs was formed in three subgroups: data of surveys and examinations; instrumental methods of research; Doppler ultrasound (27 signs in total). In accordance with the General methodology for the synthesis of hybrid fuzzy decision rules using informative features as basic variables in the interactive mode, the corresponding functions of belonging to the classes of high risk of occurrence and recurrence of brain stroke for which, using modified models E. Shortlife synthesized final predictive model.As a result of expert evaluation and mathematical modeling, it was shown that the obtained models of occurrence and recurrence of brain stroke provide confidence in the correct prognosis at the level of 0.9 and higher, depending on the amount and quality of information collected about the patient’s condition. This quality indicator was confirmed during statistical tests on representative control samples, which allows to recommend them for use in the practice of vascular surgeons and angiologists.

Keywords: astroke, cerebral membership function, the confidence in the decision being taken, the prognosis.

Full text:
KorenevskiySoavtori_3_18_1.pdf