Relational-separable models of monitoring processes at variable and unclear observation intervals

Abstract

The article is devoted to the development of combined models, methods and tools designed to solve the current problems of modeling and analysis of monitoring process data, which are repre-sented by time series and differ in variable or fuzzy observation intervals (CHRPNI). In the article, a new relational separable model (RSM) and a combined quantile algorithm are proposed to in-crease the accuracy and efficiency of modeling and analysis of the processes of CHRPNI. The rela-tional model is defined by a system of fuzzy relational relations of the first and second order ob-tained on the basis of the original sequence of data. In the combined algorithm, the results of calcu-lations obtained by SPM and models of fuzzy relational relationships were generalized with the op-timal selection of weighting factors for individual components. As a result of the conducted research by means of numerical modeling, it was established that the introduction of combined process models in the case of PNEU is rational and effective. Exam-ples of data analysis of monitoring processes of rehabilitation of diabetic patients showed certain possibilities of ensuring the accuracy of the results of the analysis of indicators and their short-term forecasting.