SELF-ORGANIZING MAPS’ APPLICATION TO CLASSIFICATION AND ANALYSIS OF SPATIALY DISTRIBUTED ENVIRONMENTAL NON-FULL DATA (PREPRINT IBRAE-1999-10)

Preprint IBRAE-99-10

Trutse A.A., Savelieva E.A., Demyanov V.V., Kanevski M.F., Timonin V.A., Chernov S.Yu. 

The work is devoted to an application of self-organizing Kohonen maps for classification of data on environmental contamination by several substancies. Values of contamination of those substancies are correlated. The reconstruction of lost data is performed on the base of obtained classification map. Detailed analysis of results of reconstruction was performed. Correspondance of statistical characteristics and spatial correlation structure of reconstructed values and initial data were analysed. 

Bibliographical reference

Trutse A.A., Savelieva E.A., Demyanov V.V., Kanevski M.F., Timonin V.A., Chernov S.Yu. SELF-ORGANIZING MAPS’ APPLICATION TO CLASSIFICATION AND ANALYSIS OF SPATIALY DISTRIBUTED ENVIRONMENTAL NON-FULL DATA. Preprint IBRAE-99-10. Moscow: Nuclear Safety Institute. November 1999. 40 p. — Refs.: 7 items.