Features and examples of clustering in the system GeoBazaDannych

Abstract

The article discusses the methodological and technical issues of the development of the GeoBazaDannych software system (GBD). The new functionality provided by the inclusion of executable data mining modules of the Wolfram Mathematica computer algebra system into the GDB is noted. In particular, it illustrates tools for preparing benchmark geodata sets for validating, testing, and evaluating related neural network models. Examples on representative data sets illustrate options for choosing the best clustering algorithms.