This article is cited in 1 paper

Anomaly recognition in recordings using fuzzy logic

S. M. Agayan^a, Sh. R. Bogoutdinov^ba, D. A. Kamaev^c, B. A. Dzeboev^a, M. N. Dobrovolsky<sup>a

a</sup> The Geophysical Center of the Russian Academy of Sciences (Moscow)
^b Schmidt Institute of Physics of the Earth of the Russian Academy of Sciences (Moscow)
^c Research and Production Association “Typhoon” (Obninsk)

Abstract: One of the areas of time series analysis is the study of their morphology. The results of such a study are used to detect various types of anomalies in the behavior of the series, moments of restructuring its behavior, etc.
This paper presents a program for exploring records using DMA methods – a new researcher-oriented data approach that makes extensive use of fuzzy mathematics. Its input data is a record expressing a process with discrete time and some property of the process whose fulfillment is of interest to the researcher.
The manifestation of a property on a record is formalized as a fuzzy structure (measure of manifestation) on the do-main of the record definition, which expresses the degree of manifestation of the property in question. The measure of manifestation of a property is the basis for dividing the record into regular, transient, and abnormal manifestations of the property in question on a record. This division gives the researcher a simple and meaningful idea of the property manifestation on a record that interests him.
The purpose of this paper is to improve on the current DMA decomposition of such a decomposition.

Keywords: anomaly recognition, time series, fuzzy logic.

UDC: 519.6

Received: 23.04.2025
Revised: 27.08.2025

DOI: 10.22405/2226-8383-2025-26-3-6-43