A data mining strategy for inductive data clustering: a synergy between self-organising neural networks and K-means clustering techniques

Published 7 November 2002

Syed Sibte Raza Abidi, Jason J. Ong

DOI10.1109/tencon.2000.888802

Citations15

TL;DR

This paper presents a technique to automate the data mining task of data clustering, i.e. to automate cluster identification/demarcation by drawing upon a synergy between the self-organising neural networks and statisticalData clustering techniques.

Abstract

Self-organising neural networks have a natural propensity to cluster well-defined data into visually distinct clusters, which can then be easily interpretable by data analysts. However, there are situations when the clustering output of the self-organising network does not render distinct clusters. In this paper, we present a technique to automate the data mining task of data clustering, i.e. to automate cluster identification/demarcation by drawing upon a synergy between the self-organising neural networks and statistical data clustering techniques. The implied hybrid of diverse data clustering techniques provides an improved strategy to (a) discover hidden similarities between data items; (b) group similar data items into distinct and well-defined clusters - i.e. with explicit boundaries between different clusters and defined cluster membership characteristics; and (c) visualise the emergent data clusters in a 2D and 3D manner. Our proposed solution is implemented in terms of a data clustering workbench (DCW) - an all-encompassing (exploratory) data mining application.

Keywords

Computer Science

Biological CyberneticsSelf-organized formation of topologically correct feature maps

9,519 Citations1982Teuvo Kohonen

In a simple network of adaptive physical elements which receives signals from a primary event space, the signal representations are automatically mapped onto a set of output responses in such a way that the responses acquire the same topological order as that of the primary events.

Proceedings of the IEEEThe self-organizing map

8,175 Citations1990Teuvo Kohonen

IEEE Transactions on Knowledge and Data EngineeringEffective data mining using neural networks

361 Citations1996Hongjun Lü, Rudy Setiono +1 more

The paper presents an approach to discover symbolic classification rules using neural networks, and demonstrates the effectiveness of the proposed approach by the experimental results on a set of standard data mining test problems.

Future Generation Computer SystemsUsing neural networks for data mining

256 Citations1997Mark Craven, Jude Shavlik

This article describes neural-network learning algorithms that are able to produce comprehensible models, and that do not require excessive training times, and argues that these methods deserve a place in the tool boxes of data-mining specialists.

Journal of Mathematical Imaging and VisionClustering properties of hierarchical self-organizing maps

192 Citations1992Jouko Lampinen, Erkki Oja

In experiments with both artificial and real data it is demonstrated that the multilayer SOM forms clusters that match better to the desired classes than do direct SOM's, classical k-means, or Isodata algorithms.

Exploratory Data Analysis by the Self-Organizing Map: Structures of Welfare and Poverty in the World

182 Citations1996Samuel Kaski, Teuvo Kohonen

Clustering Properties of Hierarchical Self-Organizing Maps

179 Citations1993Jouko Lampinen, Erkki Oja

Informatik aktuellKnowledge Extraction from Artificial Neural Networks and Applications

29 Citations1994Alfred Ultsch, Gabriela Guimarães +2 more

Self-Organising Neural Networks, as the Kohonen model, the inherent structures in high-dimensional input spaces are projected on a low dimensional space and an extension of this method from static to dynamic data is feasible.

Lecture notes in computer scienceApplying knowledge discovery to predict infectious disease epidemics

19 Citations1998Syed Sibte Raza Abidi, Alwyn Goh

It is argued that neural networks, an established computational technology, can efficaciously be used to perform predictive modelling, i.e. to explore the intrinsic dynamics of temporal data.

International Conference on Artificial IntelligenceData Mining Using Self-Organizing Kohonen Maps: A Technique for Effective Data Clustering & Visualization.

17 Citations1999Jason J. Ong, Syed Sibte Raza Abidi

The case for using neural networks as a viable data mining tool that can provide statistical insights and models from large data-sets is argued and it is demonstrated how Self-Organizing Kohonen Maps, an unsupervised learning neural network paradigm, can be efficaciously used for data mining purposes, in particular for data clustering applications.