Loading [MathJax]/extensions/MathMenu.js
A new data topology matching technique with Multilevel Interior Growing Self-Organizing Maps | IEEE Conference Publication | IEEE Xplore

A new data topology matching technique with Multilevel Interior Growing Self-Organizing Maps


Abstract:

Self-Organizing Maps (SOM) are widely used for their ability to preserve the topology in the projection. However, this topology is not perfectly preserved due to the stat...Show More

Abstract:

Self-Organizing Maps (SOM) are widely used for their ability to preserve the topology in the projection. However, this topology is not perfectly preserved due to the static structure of SOM. Therefore, we show in this paper a novel architecture of SOM which organizes itself over time. The proposed method called MIGSOM (Multilevel Interior Growing Self-Organizing Maps) is generated by a growth process which allows to adds nodes where it is necessary. The network start with a minimum number of nodes, then nodes will be added from the boundary as well as the interior of the network. The MIGSOM algorithm adds the interior nodes in a superior level of the map. As a result, the map can have three-Dimensional structure with multi-levels oriented maps. To improve the performance of the proposed algorithm, comparison of MIGSOM to the Kohonen feature Map (SOM) and the Growing Grid (GG) is made. Our experiment results demonstrate that the MIGSOM constructs better mappings than the classic SOM and GG, especially, in terms of data quantification and topology preservation.
Date of Conference: 10-13 October 2010
Date Added to IEEE Xplore: 22 November 2010
ISBN Information:

ISSN Information:

Conference Location: Istanbul, Turkey

I. Introduction

The topologies of data distributions are very important for data description. Usually, it is not easy to find a description that can give us a precise understanding of the topologies for general distributions. The Self-Organizing Map (SOM) [14] is a neural network algorithm which is based on unsupervised learning and combines the tasks of vector quantization and data projection. It is a powerful tool in several problematic domains, such as data mining/knowledge exploration, vectors quantification and data clustering tasks [6][7][11][13][15][17]. It is capable of projecting high-dimensional data onto a low dimensional space (one, two or three dimensional structure), with good neighborhood preservation between the two spaces. However, due to the static structure of SOM, the neighborhood preservation cannot always lead to perfect topology preservation. Since, SOM algorithm has needed to fix the size of the grid at the beginning of the training process. This includes the topology as well as the number of nodes. So, Ref. [19] proposes limited using data topology in SOM representation that indicates topology violations and data distribution.

References

References is not available for this document.