Industrial Automatic Control Systems and Controllers Annotation << Back Analytical and Procedural Models of Information Retrieval in Multifactor Flows with Partially Hidden Data V.M. Tyutyunnik, M.M. Al Karkhi The theory and technology of information retrieval is considered. The eight main characteristics of the information array of the Internet, which determine its uniqueness and features of the implementation of search tasks are given. Features of classical (including full-text) retrieval in a network are discussed. New concepts in the form of four defi nitions of multifactor flows (arrays) and partially hidden data, limiting the relevance and persistence of search, are introduced into the theoretical foundations of information retrieval. The main directions and features of analytical modeling of retrieval arrays, information retrieval systems and technology of search in one-factor streams of open data are presented. A random process of finding a pertinent document in a multifactorial flow with partially hidden data is represented by a cross section of random functions of two variables (vector process on the plane). It is shown that such retrieval is a process with discrete states and is analytically modeled by a discrete Markov chain (a hidden Markov model), and the finding of the most appropriate pertinent documents is performed by the Viterbi algorithm. On this basis, for the first time a procedural model for retrieval in a multifactorial array of partially hidden data is developed, which solves the problem of retrieval optimization, providing such thematic completeness of relevant data resulting from the search session by reducing the volume of the search sample, so that the semantic potential of the search term was disclosed as fully as possible, and the retrieval result was a pertinent data. Keywords: multifactorial information flow, partially hidden data, information retrieval, analytical and procedural models, relevant and pertinent data. DOI: 10.25791/asu.5.2023.1435 Pp. 08-19.