Advances and Applications in Statistics
Volume 38, Issue 2, Pages 149 - 159
(February 2014)
|
|
CLASSIFICATION OF THE COMPANIES LISTED ON THE EGYPTIAN STOCK USING SOM NEURAL NETWORK, K-MEANS AND HIERARCHICAL CLUSTERING METHODS
Rania Ahmed Hamed Mohamed
|
Abstract: This paper investigates the usefulness of the neural network, namely Kohonen’s self-organization map methodology as an alternative to the hierarchical clustering and K-means methodology in classifying the listed companies on the Egyptian Stock. In order to achieve that, 26 financial indicators have been used for 32 companies included in the investigation, published in the Statistical Yearbook issued by the Egyptian Organization for Censorship. The results indicate that Kohonen’s self-organizing map is the best tool for classification of financial data. |
Keywords and phrases: Cluster analysis, Hierarchical clustering method, K-means method, Euclidean distance, Manhattan distance, Kohonen’s self-organization maps, R-language, Matlab toolbox. |
|
Number of Downloads: 358 | Number of Views: 1175 |
|