Пожалуйста, используйте этот идентификатор, чтобы цитировать или ссылаться на этот ресурс: http://hdl.handle.net/11701/5694
Полная запись метаданных
Поле DCЗначениеЯзык
dc.contributor.authorBure, Vladimir M.-
dc.contributor.authorMitrofanova, Olga A.-
dc.date.accessioned2016-11-24T19:17:02Z-
dc.date.available2016-11-24T19:17:02Z-
dc.date.issued2016-09-
dc.identifier.citationBure V. M., Mitrofanova O. A. Prediction of the spatial distribution of ecological data using kriging and binary regression. Vestnik of Saint Petersburg University. Series 10. Applied mathematics. Computer science. Control processes, 2016, issue 3, pp. 97–105.en_GB
dc.identifier.other10.21638/11701/spbu10.2016.309-
dc.identifier.urihttp://hdl.handle.net/11701/5694-
dc.description.abstractThere are many ecological problems associated with the prediction of the spatial distribution of ecological parameters. The paper deals with one of these tasks. Suppose we have a set of ecological data measured by contact way (for example, plant leaf color intensity by N-tester), as well as an air photo of the object (for example, field). It is necessary to estimate the spatial distribution of ecological parameters. This paper proposes an approach to the solution of such problems with the joint use of kriging and binary regression. At first the uniform field areas (clusters) in the photo are determined using classification method. It is assumed that each selected area has a set of ecological data. Next, we will consider each zone separately. It is necessary to assess the level of the indicator in the given area. First variograms analysis is performed leading to the construction of the variogram model. Next construct a set of ecological parameter estimates is built using the method of ordinary kriging. Then, we set a threshold value of the ecological parameter for the zone under study. We introduced a variable that takes the value 1, if the parameter exceeds a threshold, and 0 otherwise. Thus we get a basis for logistic regression, where factors include a set of estimates predicted by kriging. In addition, these factors may include the color characteristics from air photos. As a result, we can calculate for each point the probability, if it will be close to 1, there is reason to believe that at this point the parameter value is greater than the threshold, and if the probability is close to 0, there is reason to assume that the parameter value is below the threshold. Furthermore, this paper provides an example of the approach for simulated data using R. Refs 8. Figs 4. Table 1.en_GB
dc.language.isoruen_GB
dc.publisherSt Petersburg State Universityen_GB
dc.relation.ispartofseriesVestnik of Saint Petersburg University. Series 10. Applied Mathematics. Computer Science. Control Processes;Issue 3-
dc.subjectecological dataen_GB
dc.subjectordinary krigingen_GB
dc.subjectlogistic regressionen_GB
dc.subjectRen_GB
dc.titlePrediction of the spatial distribution of ecological data using kriging and binary regressionen_GB
dc.typeArticleen_GB
Располагается в коллекциях:Issue 3

Файлы этого ресурса:
Файл Описание РазмерФормат 
09-Bure.pdf405,43 kBAdobe PDFПросмотреть/Открыть


Все ресурсы в архиве электронных ресурсов защищены авторским правом, все права сохранены.