DIAGNOSTICS OF LATENT HEMODYNAMICS BY DATA MINING: CONDITIONS OF PRACTICAL USE

Abstract

We proposed a diagnostic nomogram (map) for hemodynamic state classification, based on linear regression modeling of blood pressure (BP) parameters and Data Mining algorithm. Mapping the vector of the BP characteristics, which coordinates are the coefficients of linear regression of the systolic, diastolic on pulse pressure, the pathological changes of hemodynamics inherent in hypertension, prehypertension or clinically latent hemodynamic disturbances can be diagnosed. The normal physiology of cardiovascular interaction corresponds to the vector location within certain limits on the nomogram. We have investigated the factors influencing the accuracy of the diagnostics in a clinical setting.