Multi-level Model for Structuring Heterogeneous Biomedical Data in the Tasks of Socially Significant Diseases Risk Evaluation

Abstract

The article provides an overview of modern approaches to building multi-level models for structuring heterogeneous data of different nature. A generalized multilevel model for structuring heterogeneous data is proposed and its adaptation for processing biomedical data in the tasks of socially significant diseases risk evaluation is described by the example of risk evaluation of congenital malformations among children from areas in the Bryansk region radioactively contaminated after the Chernobyl disaster. The developed model allows for enriching the available data, taking into account more factors and hidden relationships between them, facilitating intelligent analysis and enhancing the quality of its results. As a result of the study, no statistically significant excess of the frequency of anencephaly, hydrocephalus and encephalocele in children in more radiation-contaminated southwestern territories was found compared to the average regional data. The results obtained presumably indicate the influence of the radiation factor on the increased incidence of microcephaly in southwestern territories relative to the average regional values without southwestern territories for a sixteen-year period (1999–2014).