Practice of applying models of clustering of incoming semi-structured data in management systems

Abstract

A scheme for solving the problem of binary clustering of incoming data in the control system is proposed. Different ways of presenting the initial data of the clustering problem are considered. Methods of sequential shortening and sequential merging of clusters and the model of initial placement of clusters are considered. Estimating the number of clusters is given to solve the clustering problem. A binary clustering method has been proposed for points in a circle. The clustering practice was performed based on the research results, and the developed model and algorithms were implemented. The optimal variant of the processing process based on clustering methods for semi-formed information sources is proposed based on the proposed structure.