G.A. Mashevskiy - Ph. D. (Eng.), Associate Professor, Department of Biotechnical Systems, Saint-Petersburg Electrotechnical University «LETI»
M.V. Semenovskaya - Under-graduate Student, FSBI Research Institute of Obstetrics and Gynecology them D.O. Ott (St.-Petersburg)
The article concerns the question of modern information technology application for the efficiency of diagnosing the presence of patients intoxication with hydrosulfide ion or ferrous iron improvement. For these purpose there was developed the Bayesian network model of intoxication form decision, based on electrocemical model of Ag2S- and Pt-electrodes reaction in a presence of HS- and Fe2+ ions. There were selected and justified nodes of the network, as well as established a priori and conditional probabilities of states of these nodes. After insertion in the network the data from anamnesis and the results of the patients urine electrochemical analysis the user receives the calculated probabilities of the presence of some form of intoxication.
Experimental testing showed a high value of the proposed model. At the same time the carried out researches showed that the accuracy of the model depends on the area of the electrode potential In particular it remains a controversial question of selecting the threshold values of the potential of a platinum electrode for Fe2+-intoxication diagnostics purposes. In addition, the proposed version of the model does not allow to make an unambiguous conclusion about the nature of the patient\'s pathology in case if the potential of Ag2S- and Pt-electrodes at the same time are in positive values. Thus there remains a need model refinement. Nevertheless, the problems noted above show themselves rarely enough and designed network can be used to facilitate the setting process of the diagnosis of the presence of a patient intoxication, included in the information system for patient water and salt exchange monitoring.