A novel clustering method based on spatial operations

Authors:
Hui Wang
Affiliations:
School of Computing and Mathematics, University of Ulster at Jordanstown, Newtownabbey, Northern Ireland, UK
Venue:
BNCOD'06 Proceedings of the 23rd British National Conference on Databases, conference on Flexible and Efficient Information Handling
Year:
2006

Citing 8
Cited 1

Algorithms for clustering data

Algorithms for clustering data
Clustering related tuples in databases

The Computer Journal
Incremental Clustering for Mining in a Data Warehousing Environment

VLDB '98 Proceedings of the 24rd International Conference on Very Large Data Bases
Clustering Categorical Data: An Approach Based on Dynamical Systems

VLDB '98 Proceedings of the 24rd International Conference on Very Large Data Bases
CCAIIA: Clustering Categorial Attributed into Interseting Accociation Rules

PAKDD '98 Proceedings of the Second Pacific-Asia Conference on Research and Development in Knowledge Discovery and Data Mining
STING: A Statistical Information Grid Approach to Spatial Data Mining

VLDB '97 Proceedings of the 23rd International Conference on Very Large Data Bases
Mining Market Basket Data Using Share Measures and Characterized Itemsets

PAKDD '98 Proceedings of the Second Pacific-Asia Conference on Research and Development in Knowledge Discovery and Data Mining
Grid-Clustering: An Efficient Hierarchical Clustering Method for Very Large Data Sets

ICPR '96 Proceedings of the 13th International Conference on Pattern Recognition - Volume 2

Lattice Machine Classification based on Contextual Probability

Fundamenta Informaticae - To Andrzej Skowron on His 70th Birthday

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper we present a novel clustering method that can deal with both numerical and categorical data with a novel clustering objective and without the need of a user specified parameter. Our approach is based on an extension of database relation – hyperrelations. A hyperrelation is a set of hypertuples, which are vectors of sets. In this paper we show that hyperrelations can be exploited to develop a new method for clustering both numerical and categorical data. This method merges hypertuples pairwise in the direction of increasing the density of hypertuples. This process is fully automatic in the sense that no parameter is needed from users. Initial experiments with artificial and real-world data showed this novel approach is promising.