FM: A fast map algorithm for data clustering

被引：0

作者：

Wang, L ^{[1
]}

Wang, ZO ^{[1
]}

机构：

[1] Tianjin Univ, Inst Syst Engn, Tianjin 300072, Peoples R China

来源：

2002 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS, VOLS 1-4, PROCEEDINGS | 2002年

关键词：

clustering; data mining; neural networks; self-organizing feature maps;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We propose a fast map (FM) algorithm, which has significant advantages for knowledge discovery applications due to its low running time and hierarchical clustering capability compared with the exiting similar algorithms. The FM algorithm is presented in detail and the effect of a spread factor is investigated. The spread factor can control the growth of network structure (number of nodes and connections), and it is also presented as a method of achieving hierarchical clustering of a data set. Only a small network is created at the beginning with a low spread factor, further analysis is conducted on selected sections of the data, which have smaller volume. Therefore, this method facilitates the analysis of even very large data sets.

引用

页码：55 / 59

页数：5

共 6 条

[1] Dynamic self-organizing maps with controlled growth for knowledge discovery
Alahakoon, D
Halgamuge, SK
Srinivasan, B
[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS, 2000, 11 (03): : 601 - 614
[2] [Anonymous], DATA MINING CONCEPTS
[3] Blackmore J., 1995, THESIS U TEXAS AUSTI
[4] GROWING CELL STRUCTURES - A SELF-ORGANIZING NETWORK FOR UNSUPERVISED AND SUPERVISED LEARNING
FRITZKE, B
[J]. NEURAL NETWORKS, 1994, 7 (09) : 1441 - 1460
[5] Kohonen T., 1997, Self-organizing Maps, V2nd ed.
[6] MARTINETZ T, 1991, NEURAL GAS NETWORK L, P397

← 1 →