计算机工程与应用 ›› 2013, Vol. 49 ›› Issue (17): 112-115.

• 数据库、数据挖掘、机器学习 • 上一篇    下一篇

改进的基于SOM的高维数据可视化算法

王志省,许晓兵   

  1. 上海理工大学 管理学院 管理科学与工程系,上海 200093
  • 出版日期:2013-09-01 发布日期:2013-09-13

Improved SOM-based high-dimensional data visualization algorithm

WANG Zhisheng, XU Xiaobing   

  1. Department of Management Science and Engineering, School of Business, University of Shanghai Science and Technology, Shanghai 200093, China
  • Online:2013-09-01 Published:2013-09-13

摘要: 通过对高维数据可视化方法的系统研究,提出了一种新的基于自组织映射(Self-Organizing Map,SOM)的算法。为了表现该方法的特点,将其称为三维自组织映射(Three-Dimensional SOM,TDSOM)。它在对高维数据记录集进行SOM分析后将其投影到三维坐标系中的特定的点集上,最终形成三维模型。该模型弥补了传统模型难以清晰准确地展现高维数据的缺陷,并且新模型着重于在一个比二维平面更为广阔的三维立体空间中展现海量数据。使用者通常可以根据当前领域的专业知识在分析模型的基础上得出有意义的模式。新方法可以广泛使用在数据挖掘和模式识别等领域。

关键词: 自组织映射, 神经网络, 高维数据可视化, 聚类分析

Abstract: A new high-dimensional data visualization algorithm based on the Self-Organizing Map(SOM) is demonstrated. The algorithm is named TDSOM(Three-Dimensional Self-Organizing Map) to describe its special characteristics. After being trained with SOM network, high-dimensional data is projected into particular point sets in the three-dimensional coordinate system. A three-dimensional model is created by the algorithm. Through the experiment, TDSOM is proved to be much more accurate and more analytical than the traditional methods in displaying the high-dimensional data. The main innovation of the new TDSOM algorithm is the presentation of the large data in three-dimensional coordinate system which is much vaster than the two-dimensional one. What’s more, users are able to discover some interesting patterns according to their own areas through the model. The method can be widely applied in areas such as data mining, pattern recognition and so on.

Key words: self-organizing map, neural networks, data visualization, clustering analysis