Approach to clustering objects

Igor A. Botygin, Sergey G. Kataev, Valeriy A. Tartakovskiy, Anna I. Sherstneva

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)


Relevance of the work is due to the need to develop universal information-analytical approaches to extract knowledge from the rapidly growing volume of geophysical data. One of the main problems in the processing of geophysical data is to find in it objectively existing laws which could become the basis for diverse, including forward-looking, behavior models of selected parameters of geophysical fields. And that data clustering technologies are the foundation for the software development of similar information systems analysis of unstructured data. The main aim of the study is to develop a method of experimental data clustering of geophysical nature on the basis of allocation of structures for solving problems of the analysis of unstructured information when studying and controlling complex systems. The methods used in the study: classical and modern methods and clustering algorithms, graph theory algorithms, test case of clustering of a geophysical field of meteorological parameters from the territory of the northern part of Eurasia. The results. The authors developed a new algorithm of structures allocation in the initial geophysical field, which allows decomposing the test space into fields with the same behavior of the studied parameters based on spatial characteristics. The algorithm is based on the structuring of the various expansions of geophysical fields (season, anomaly, etc.) and provides a wide range of information on the object in the form of sets of parameters of the selected structures. This information, along with the accompanying empirical relationship between the parameters is considered as a generalization of the experimental characterization of the object and is the basis for the formation of hypotheses and behavior models. In addition, a structural model of the space of a meteorological parameter provides the ability to compress primary data without significant loss of semantic value of the target geophysical field.

Original languageEnglish
Pages (from-to)78-85
Number of pages8
JournalBulletin of the Tomsk Polytechnic University, Geo Assets Engineering
Issue number12
Publication statusPublished - 2015


  • Clustering
  • Geophysical field
  • Graph theory
  • Method of allocation of structures
  • Structure
  • Time series
  • Weather observations

ASJC Scopus subject areas

  • Materials Science (miscellaneous)
  • Fuel Technology
  • Geotechnical Engineering and Engineering Geology
  • Waste Management and Disposal
  • Economic Geology
  • Management, Monitoring, Policy and Law

Fingerprint Dive into the research topics of 'Approach to clustering objects'. Together they form a unique fingerprint.

Cite this