Showing papers by "Yannis Theodoridis published in 2002"

Revisiting R-tree construction principles

Abstract: Spatial indexing is a well researched field that benefited computer science with many outstanding results. Our effort in this paper can be seen as revisiting some outstanding contributions to spatial indexing, questioning some paradigms, and designing an access method with globally improved performance characteristics. In particular, we argue that dynamic R-tree construction is a typical clustering problem which can be addressed by incorporating existing clustering algorithms. As a working example, we adopt the well-known k-means algorithm. Further, we study the effect of relaxing the two-way split procedure and propose a multi-way split, which inherently is supported by clustering techniques. We compare our clustering approach to two prominent examples of spatial access methods, the R- and the R*-tree.

Revisiting R-Tree Construction Principles

Abstract: Spatial indexing is a well researched field that benefited computer science with many outstanding results. Our effort in this paper can be seen as revisiting some outstanding contributions to spatial indexing, questioning some paradigms, and designing an access method with globally improved performance characteristics. In particular, we argue that dynamic R-tree construction is a typical clustering problem which can be addressed by incorporating existing clustering algorithms. As a working example, we adopt the well-known k-means algorithm. Further, we study the effect of relaxing the "two-way" split procedure and propose a "multi-way" split, which inherently is supported by clustering techniques. We compare our clustering approach to two prominent examples of spatial access methods, the R- and the R*-tree.

An efficient and effective algorithm for density biased sampling

Abstract: In this paper we describe a new density-biased sampling algorithm. It exploits spatial indexes and the local density information they preserve, to provide improved quality of sampling result and fast access to elements of the dataset. It attains improved sampling quality, with respect to factors like skew, noise or dimensionality. Moreover, it has the advantage of efficiently handling dynamic updates, and it requires low execution times. The performance of the proposed method is examined experimentally. The comparative results illustrate its superiority over existing methods.

A Survey on Pattern Application Domains and Pattern Management Approaches

Abstract: Data intensive applications produce complex information that is posing requirements for novel Database Management Systems (DBMSs). Such information is characterized by its huge volume of data and by its diversity and complexity, since the data processing methods such as pattern recognition, data mining and knowledge extraction result in knowledge artifacts like clusters, association rules, decision trees and others. These artifacts that we call patterns need to be stored and retrieved efficiently. In order to accomplish this we have to express them within a formalism and a language. In this report we review the concept of patterns and their applicability in several research domains related with the proposed work and we define the knowledge domain related with the PANDA project. It is important to interrelate these domains in order to be able to define the problem in comprehensive and complete way and come up with requirements on how a management system for patterns should be. We examine the different types of patterns that are extracted from a data set, in order to gather the necessary requirements for the definition of a pattern model. This model constitutes the heart of the Pattern Base Management System that will be designed.