N.A. Korenevskiy - Dr.Sc. (Eng.), Professor, Southwest State University
I.A. Klyuchikov - Dr.Sc. (Eng.), Professor, Southwest State University
V.I. Serebrovskiy - Dr.Sc. (Eng.), Professor, Kursk State Agricultural Academy n.a. professor I.I. Ivanov
E.B. Ryabkova - Post-graduate Student, Southwest State University

Rank analysis method is developed to simplify computational procedures while maintaining a high informative prognostic and diagnostic features. This method can be implemented by means of modern microprocessors and microcontrollers. Locking mechanism of the switching points of the physiological signal is as the basis for rank analysis method. Next these points are sorted by special the procedure. Ranging signal parameters on the switching point is based on the allocation of the processes studied a set of indicators (periodometrical, frequency, amplitude, asymmetric, etc.). Different time components are calculated with respect to the signal switching points and ranked in descending speed of their changes. As the switching points of the inflection point, violations of monotonicity, minima, maxima, tears, etc can be selected. Grades are determined and ordered by the number "fast" components. Ranking ends either exhaustion switching points at the next step of ranking, or after going beyond the informativeness studied parameters. The period between two adjacent switching (periodometrical characteristic) can be selected as the speed the process characteristics and its ranks. Different areas of the multidimensional feature space correspond to different states of man. This allows the use of the theory of pattern recognition in its unfuzzy and fuzzy interpretation to solve classification problems. One of the major drawbacks of the proposed method is the treatment of electrophysiological signals as its high sensitivity to high-frequency and low frequency interferences, which requires a high removal rate and systems pretreatment baseline information. Currently, this problem is relatively easy solved by integrated chips for collecting and pre-processing the analog signal (chip type AFE). AFE chips are produced specifically for the removal and pretreatment electrocardiological signals by Texas Instruments. These chips can make amplification, filtering, analog-to-digital conversion and forming of standard protocols such as SPI. This allows the AFE chips connect directly to high-performance microcontrollers and signal processors. Performance of this method selection of informative features tested on tasks assessing the severity of diffuse toxic goiter. Technique was used with the inclusion of a continuous recording pulsogram patient during sequential execution antiorthostatic sample on a special turntable. Three classes of severity of the disease has been allocated at the expert level. Checking the quality of classification in the control samples of 50 people for each class showed that the diagnostic efficiency of the classification rules obtained is 0.87.