论文信息 - INVESTIGATION OF THE QUALITY OF MAPPING VECTORS OBTAINED BY QUANTIZATION METHODS

INVESTIGATION OF THE QUALITY OF MAPPING VECTORS OBTAINED BY QUANTIZATION METHODS

In this paper, the quality of mapping vectors, obtained by vector quantization methods (self- organizing map and neural gas), is investigated. Multidimensional scaling is used for mapping. Two crite- ria (topology preserving measure and Spearman's rho) are applied for the estimation of mapping quality.

Olga Kurasova | Alma Molyt | O. Kurasova | Alma Molyt

[1] Gintautas Dzemyda,et al. Optimal decisions in combining the SOM with nonlinear projection methods , 2006, Eur. J. Oper. Res..

[2] P. Groenen,et al. Modern multidimensional scaling , 1996 .

[3] Pablo A. Estévez,et al. Cross-entropy embedding of high-dimensional data using the neural gas model , 2005, Neural Networks.

[4] A. Konig. Interactive visualization and analysis of hierarchical neural projections for data mining , 2000 .

[5] Gintautas Dzemyda,et al. Heuristic approach for minimizing the projection error in the integrated mapping , 2006, Eur. J. Oper. Res..