ZASTOSOWANIE SAMOORGANIZUJACYCH SIECI NEURONOWYCH KOHONENA SAMOORGANIZUJACYCH SIECI NEURONOWYCH KOHONENA (REJON UJKOWICE - BATYCZE)

摘要

This paper presents the application of Kohonen's Self Organizing Networks in classification of seismic waveform. The classification is one of the basic elements of seismofacies analysis andit often leads to significant exploratory conclusions. Important elements of this kind of analysis are:|selection of seismic attributes and usage ofappropriateclustering method. There were used AVA attrilbutes,whichincludeinformationaboutpetrophysicalproperties of rocks. There used two additionalimulti-dimensional methods to examine seismic facies distribution on selected area: classification chosen crossplot intercept-gradient area and classification earned out by method which minimizes the:product of objects distances in groups. Verification of optimal method for data classification wasmade based on observation of clusters shape and their characteristic due to insufficient informationfrom well.