cNnet-SCHEME OF THE MARKOV ALGORITHM

The homology of the N-schema of the Markov algorithm is subjected to comprehensive scientific analysis. From the composition of the homology of the algorithm, new variants of modifying its schemes are distinguished. The topos of hashes is introduced into the categorical сN-scheme of the Markov algorithm and the theoretical provisions of this scheme of the algorithm are developed in order to establish new, previously unknown properties of the Markov algorithm. The hash topos stores data on all interactions of words, which is used to check the word received from the input data sample for belonging to its chain of Markov occurrences. The definition of belonging to one's chain of occurrences is used to classify cycles, otherwise called λ-voids, which is necessary to overcome the looping of algorithms on data from the input sample. The economic interconnections of organizational systems are investigated, where these interconnections rely on morphisms, and words in Markov alphabets are objects of these systems. Economic relationships in organizational systems are subjected to scientific analysis to solve the problem of synthesizing a problem-oriented management system for combating money laundering and terrorist financing. Additionally, several morphisms are introduced, in particular, those that are responsible for weighting values, their theoretical justification is given, as a practical result of their use, the implementation of weighting mechanisms in Markov networks is proposed. A сNnet-scheme of the Markov algorithm with weighting is synthesized for application in Markov networks. Several weighting methods are shown: cloned source, types of morphisms, word morphisms, taking into account both the associator and the induction. It has been established that the use of hashing in the сNnet-scheme of the Markov algorithm provides Markov networks with sufficiently economical memory consumption for such a significant and resource-intensive procedure as the detection and classification of λ-voids. It is concluded that the Markov network has acquired internal, auxiliary memory, which stores the entire topos, which significantly expands its potential for processing both words and morphisms from input data samples from a variety of data sources.