Methodological aspects of natural territories zoning using machine learning

D.O. Krivoguz, R.V. Borovskaya

Azov Sea Research Fisheries Institute (FSBSI “AzNIIRKH”), Kerch, Sverdlova St., 2

E-mail: krivoguz_d_o@azniirkh.ru

DOI: 10.33075/2220-5861-2020-1-14-20

UDC 544.623

Abstract:

     The article discusses a modern approach to natural territories zoning by machine learning method. Zoning of the territory is based on an integrated assessment of the quality of environmental indicators, taking into account the maximum possible number of factors that can fully describe the properties and features of the analyzed territory. The purpose of the work is to form a modern mathematical methodology based on machine learning methods, capable of significantly improving modern approaches of any territory zoning. The article analyzes current trends in zoning for various territories.

     Authors have analyzed the zoning algorithm using clustering, and also highlighted its main stages, which include obtaining and normalizing data, determining the propensity of a dataset to be divided into sets, identifying the optimal number of clusters and zoning. It was concluded that the effectiveness of any clustering algorithm is determined by its achievement of the compactness hypothesis, which consists in the fact that similar objects are much more likely to be in the same class than in different ones. The authors highlighted both the strengths and weaknesses of this approach. The authors consider such qualities as objectivity, accuracy, simplicity of modifiability and settings as positive ones while the strong dependence on the quality of the data of carried out standardization is considered to be negative, which, with any significant deviations in these aspects, can lead to a significant distortion of the results.

Keywords: territories zoning, fishery zoning, machine learning, environmental problems, clustering, data normalization.

To quote, follow the DOI link and use the Actions-Cite option or copy:

[IEEE] D. O. Krivoguz and R. V. Borovskaya, “Methodological aspects of natural territories zoning using machine learning,” Monitoring systems of environment, no. 1, pp. 13–20, Mar. 2020.

LIST OF REFERENCES

  1. Krivoguz D.O., Zaharova Ju.B. Geospatial analysis of ecology-economic growth of Kerch Peninsula using GIS // Geoinformatika. 2018. № 1. pp. 52–55.
  2. Bocharnikov V. Ecological and geographical mapping of Russian economic regions based on GIS technologies // Journal of Volgograd State University. Series 3. Economics. Ecology. 2016. № 3 (3). pp. 163–176.
  3. Matishov G.G. The Sea of Azov: Recent abrasion processes and problems of coastal protection / G.G. Matishov, L.A. Bespalova L.A., O.V. Ivlieva [et al.] // Doklady Earth Sciences. 2016. № 2 (471). pp. 1269–1272.
  4. Vojtov I.V. Scientific-innovational principles of geoecological zoning of administrative territories of Belarus / I.V. Vojtov, M.A. Gatih, L.S. Lis [et al.] // Bulletin of the BelarusianRussian University. 2009. № 22 (1). pp. 113–127.
  5. Mihajlova A.A., Vagizov M.R. Technique for processing data of Earth remote sensing with the use of information technologies and allometric correlations for definition of forest stands indicators // Advances in Current Natural Sciences. 2018. № 4. pp. 80–85.
  6. Kondrik D., Kazakov E., Pozdnyakov D. A synthetic satellite dataset of the spatio-temporal distributions of Emiliania huxleyi blooms and their impacts on Arctic and sub-Arctic marine environments (1998–2016) // Earth System Science Data. 2019. № 1 (11). pp. 119–128.
  7. Demchev D. Sea ice drift tracking from sequential SAR images using accelerated-KAZE features / D. Demchev, V. Volkov, E. Kazakov [et al.] // IEEE Transactions on Geoscience and Remote Sensing. 2017. № 9 (55). pp. 5174–5184.
  8. Ruzhnikova N.N. Geoeсological zoning of the White Sea in the transport of oil products // Bulletin of Higher Educational Institutions. North Caucasus Region. Natural Sciences. 2012. № 6. pp. 94–98.
  9. Denisov V.V., Il’in G.V. Water areas zoning as implement of nature management optimization in connection with oil and gas resources development on the Arctic shelf // Arctic and Antarctic Research. 2008. № 2 (79). pp. 134–144.
  10. Matishov G.G., Balykin P.A., Ponomareva E.N. Fishery zoning is the first stage of spatial planning of marine activities in the Arctic // Science in the South of Russia. 2018. № 2 (14). pp. 33–41
  11. Semenova A.Ju. Socio-economic priorities for preserving and promoting the health of the population of the Republic of Crimea in the context of innovative development of the region // Economics and Management: Theory and Practice. 2018. № 3 (4). pp. 49–54.
  12. Shi T., Horvath S. Unsupervised learning with random forest predictors // Journal of Computational and Graphical Statistics. 2006. № 1 (15). pp. 118–138.
  13. Kachanovskij Ju.P., Korotkov E.A. Preprocessing data for training of neural networks // Fundamental Research. 2011. № 1 (12). pp. 117–120.
  14. Aydın A., Eker R. Fuzzy rule-based landslide susceptibility mapping in Yığılca Forest District (Northwest of Turkey) // Journal of the Faculty of Forestry Istanbul University. 2016. № 66 (2). pp. 559–571.
  15. Mirzaei J. Assessment of land cover changes using RS and GIS (case study: Zagros forests, Iran) / J. Mirzaei, A. Mohamadi, Z. Heidarizadi [et al.] // Journal of Materials and Environmental Science. 2015. № 6 (9). pp. 2565–2572.
  16. Campello R. Density‐based clustering / R. Campello, P. Kroger, J. Sander [et al.] // Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery. 2019. № 2 (10). e1343.
  17. Sivogolovko E., Thalheim B. Semantic approach to cluster validity notion // Advances in Databases and Information Systems. 2013. Vol. 186. pp. 229–239.
  18. Sivogolovko E.V. Methods of evaluation of clusterization quality // Computer Tools in Education. 2011. № 4. pp. 14–31.

Loading