Application of System Diversity to Processing Unstable Data

The article examines the effectiveness of system diversity as a means of increasing the accuracy of processing unstable data obtained from heterogeneous sources through various AI techniques. In order to carry out the research, an analytical model was created, through which a quantitative assessment of the influence of diversity is obtained. In order to achieve an adequate assessment, the concept of diversity coefficient is used, postulated on two types of errors - detectable and undetectable by comparison.