Graficzna ocena jakości odwzorowania struktury grupowej na mapie Kohonena
Graphical quality assessment of group structure mapping on the Kohonen’s map

Summary/Abstract: In the article the authors used generated data sets with known spatial structure of units. For each of the data set there was built the neural network type of SOM (Self Organiz-ing Map). For each neural network there was made a visualization of coefficients of quality mapping. This approach allows to decide which part of Kohonen’s map is responsible for mapping errors as well as to recognize those units and clusters which are correctly and in-correctly grouped. The authors conducted a local assessment of the quality of clustering for each unit.

• Details
• Contents