Prediction of Coronary Risk Using a Multichannel System with Redundant Decisions and Associative Choice

The article describes a method for making redundant decisions with subsequent associative choice between them. The method is intended for use in an automated system for predicting coronary heart disease. It can be implemented in mathematical models for determining coronary risks based on three groups of informative features; algorithmic support for the synthesis of neural network models and for fuzzy inference models; an associative decision-making algorithm for synthesizing redundant channel classifiers; a redundant channel classifier providing associative choice between redundant decisions. Experimental studies have shown that the diagnostic efficiency of models with associative choice exceeds that of currently used models by 10-16%, while the use of associative choice and redundant channels makes it possible to increase diagnostic efficiency by 8% compared to the diagnostic efficiency of individual channels.