Zastosowanie jednowymiarowej sieci SOM do wyboru cech zmiennych w grupowaniu dynamicznym
Applying the one-dimensional SOM network to select variables in dynamic clustering

Summary/Abstract: In the article the author proposes an original procedure for selecting the features in dynamic clustering. The author verifies the potential of the dynamic clustering method, such as: self-learning neural network type of GNG. To speed up and improve the results of the clustering, the author reduces the number of variables using SOM network. The simulation experiment shows the effectiveness of this approach. This approach allows a considerable speed up of the process of dynamic clustering without losing the quality of clustering.

• Details
• Contents