A growing self-organizing algorithm for dynamic clustering

被引：0

作者：

Ohta, R ^{[1
]}

Saito, T ^{[1
]}

机构：

[1] Hosei Univ, Dept Elect & Elect Engn, Tokyo 142, Japan

来源：

IJCNN'01: INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-4, PROCEEDINGS | 2001年

关键词：

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents a novel growing self-organizing algorithm for dynamic clustering. Controlling a signal counter of each cell, the network can grow. Also, if there exists undesired cell for the clustering, the cell can be deleted virtually. Our algorithm can reinforce the clustering function and the learned network can adapt flexibly to time-variant input space.

引用

页码：469 / 473

页数：5

共 6 条

[1] Dynamic self-organizing maps with controlled growth for knowledge discovery [J].