Conference publications

Abstracts

XXII conference

Data analysis and forecast using the class of logical decision functions

Polyakova G.L.

NGU; 12 Pirogova avenue, Novosibirsk, 630090, Russia

In this paper the method of analysis of empirical information represented in the form of multidimensional time series is suggested. Results of statistical analysis are represented in the form of logical rules reflected cause-effect relations of object under investigation. The logical-and-probability model can to be presented in the form of a decision tree. The regularity list corresponds to a partitioning of the initial space of variables on a number of areas; each one corresponds to the true of regularity from the list. Effectiveness of the suggested method is shown by solving applied problems in the sphere of ecology

• view as PDF (in Russian) (PDF)