This article gives a short overview of existing and most known clustering methods both based on conversional algorithms and connected with Artificial Intelligence as well. It deals with the definition of clustering and provides a classification of methods as k- means, c-means, graph theory. Moreover, their basics are described. In addition, the mentioned ways of clustering analysis are compared and for each the advantages and disadvantages are headlined. Comparison table of operational complexity and algorithms’ data format is presented. The article covers the use of an advanced algorithms based on the neural network application as Self-Organizing Kohonen network, as an unsupervised learning method, and Probability Neural Network, a supervised learning one, for cluster analysis. The drawbacks and the results of artificial clustering are provided as well.The software Matlab 7.0 is used to represent the implementation and model the methods.All pictures below, except the fig.1, is done on Matlab 7.0.
Ця робота ліцензується відповідно до Creative Commons Attribution 4.0 International License.
| Переглядів: 0 | Завантажень: 0 |