HUMAN BIOLOGICAL AGE: REGRESSION AND NEURAL NETWORK MODELS

The purpose of this study is to develop methods for determining the biological age (BA) and pace human aging based on anthropometric biochemical laboratory biomarkers, comparing accuracy BA determination using regression neural network analysis. In 735 practically healthy people aged from 20 79, we determined blood plasma concentrations glucose insulin serum total cholesterol, cholesterol high, low very-low density lipoproteins, triglycerides, urea, creatinine, transaminases, alkaline phosphatase. Also, conducted measurements a standard oral tolerance test calculated HOMA index. Age recognition was carried out multiple equation, which connects examinees with their parameters, allows calculate metabolic person an absolute error 6.92 years. This sufficient reveal at risk accelerated aging. use algorithm deep learning determine 4.57 years, distinguish between physiological increases person’s age, its reduced by 40%.