Weekly seminar on November 25, 2017, 10.00 will feature a report by PHD, Mirzazade Irada Hatam gizi, department Senior scientific researcher of “Applied mathematics”, entitled “Development of intellectual-information system for diagnostics and monitoring of toxic gas poison”.
This thesis is dedicated to development of an intelligent information system of diagnosis and monitoring of poisoning by toxic substances with the use of carbon monoxide as an example. Cases of carbon monoxide poisoning have recently become more frequent in domestic life, at work etc. The cases of carbon monoxide poisoning occupy the third place by their incidence and occurrence after cardiovascular and oncological diseases.
The application of information technologies is due to complexity, indetermination and a great quantity of parameters in the problem under consideration. A class of toxic substances is revealed which cause symptoms similar to those of carbon monoxide poisoning thus bringing confusion and difficulties to diagnosing poisoning by the latter. Carbon monoxide poisoning has severe consequences, after some time a patient affected by carbon monoxide develops neurosomatic and cardiovascular diseases, cardiac muscle injury may lead to myocardial infarction and even death. For this reason construction of an intelligent information system is proposed with diagnosis and monitoring of the state of carbon monoxide poisoning.
On the basis of classification of bio cybernetic systems for diagnosis two approaches are suggested in this research work: a simple probabilistic and very intricate probabilistically determinated one. On the strength of this Bayes’s method and neuronal networks are applied for diagnosis of carbon monoxide poisoning.
This work proposes a time series method for monitoring the state of a patient after treatment of carbon monoxide poisoning. The said method allows to trace dynamics of indices in time intervals and detect a more important index for observation of treatment resistant symptoms and elimination of excessive checks. For comparison of the indices in time intervals parametric and non-parametric criteria of biostatistics are employed.
In this thesis architecture is elaborated and the intelligent information system is made for which a software package is prepared, numerous experiments have been carried out.