Finding Structurally Similar Objects Based on Data Sorting Methods

P. 826–835.

Myachin A. L.

We present methods for finding patterns using sorting algorithms. Two modifications of diffusion-invariant pattern clustering are proposed, which allow finding structurally similar objects under study. The proposed modifications allow you to reduce the time spent on searching for templates, which allows you to work with big datasets. It assumes endogenous quantification and distribution of income patterns. The methodology for adjusting the obtained results is described. The method for searching structurally close objects in the presence of errors in the initial data selection is suggested. The possibility of correcting the final results is very important given the high sensitivity of the methodology to the presence of errors in the initial dataset. The proposed algorithmic solutions are demonstrated using a practical example. The results are compared with varying methods of cluster analysis.

Keywords: pattern analysis, diffusion-invariant pattern clustering, Sorting algorithms

Publication based on the results of: Study of models and methods of decision making under conditions of deep uncertainty (2022)

In book

Intelligent Computing: Proceedings of the 2022 Computing Conference, Volume 1

Springer, 2022.