Using Force-Based Graph Layout for Clustering of Relational Data

摘要

Data clustering is essential problem in database technology - successful solutions in this field provide data storing and accessing optimizations, which yield better performance characteristics. Another advantage of clustering is in relation with ability to distinguish similar data patterns and semantically interconnected entities. This in turn is very valuable for data mining and knowledge discovery activities. Although many general clustering strategies and algorithms were developed in past years, this search is still far from end, as there are many potential implementation fields, each stating its own unique requirements. This paper describes data clustering based on original spatial partitioning of force-based graph layout, which provides natural way for data organization in relational databases. Practical usage of developed approach is demonstrated.