AN APPROACH TO FUNCTIONAL MODELING OF MARKOV NETWORKS. Part 2

The study of the theoretical-algorithmic approach to functional modeling of Markov networks, which began in the first part of the study, continues. The idea of extracting models of objects and systems of the world around us from input sample sets by processing input data samples with a cNnet-scheme of the Markov algorithm is considered. Additionally and in parallel, a model of the Markov network itself, formed from these algorithm schemes, is synthesized. Technically, the Markov network model is the result of the formation (accumulation and filling with processed data) of four trees Tree4 from the cNnet-scheme of the Markov algorithm, shown in the first part of the study. From a general theoretical perspective, the possibilities of homology of Markov algorithms in the problem of modeling objects and systems of the world around us are explored. The main goal of the study of such models is the synthesis of new, previously non-existent systems and objects of the world around us, taking into account the knowledge extracted, processed and generalized about previously known objects of the world around us. We consider part of this homology algorithm with output that cannot be considered Markov normal. It has been determined that during the self-synthesis of models of both known and new objects and systems of the world around us, it is necessary to develop and develop methods and methods for assessing the results of modeling. In particular, the ability of a person to adequately and fully perceive (process) the results of modeling both known and new interacting objects is of scientific interest. In the course of theoretical considerations, it was established that the requirements for the adequacy and completeness of processing modeling results must be presented both to mathematical compositions and to artificial intelligence that receives data from these compositions. It is shown that a categorical language can be used to exchange data between mathematical compositions and artificial intelligence. It has been established that a person’s ability to interpret categorical language is limited, mathematical compositions do not have such potential limitations, and artificial intelligence will have them if it follows the algorithm’s instructions to recode categorical language into Aristotelian categories.