This thesis describes a new approach of relational database design using the SET conceptual model. The SET conceptual model is used for information modelling. The database schema generated from the information modelling is called the SET schema. The SET schema consists of the declarations of all the sets of the database schema. A domain graph can be constructed based on the information declared in the SET schema. A domain graph is a directed graph with nodes labelled with declared sets and arcs labelled with degree information. Each are in the domain graph points to a node S from a node labelled with an immediate domain predecessor of S. The new method of table design for the relational database involves partitioning the domain graph into mutually exclusive <1,1>-connected components based on the degree information. These components (subgraphs) are then transformed into tree structures. These trees are extended to include the domain predecessors of their nodes to make them predecessor total. The projections of these extended trees into the value sets labelling their leaf nodes form a set of relations which can be represented by tables. This table design method is described and presented in this thesis, along with d program that automates the method. Given a schema of the SET model, together with some degree information about defined sets that a user must calculate based on the intention of the defined sets, the program produces a relational database schema that will record data for the SET schema correctly and completely. / Science, Faculty of / Computer Science, Department of / Graduate

In the field of machine learning, methods for learning from single-table data have received much more attention than those for learning from multi-table, or relational data, which are generally more computationally complex. However, a significant amount of the world's data is relational. This indicates a need for algorithms that can operate efficiently on relational data and exploit the larger body of work produced in the area of single-table techniques. This thesis presents algorithms for learning from relational data that mitigate, to some extent, the complexity normally associated with such learning. All algorithms in this thesis are based on the generation of random relational rules. The assumption is that random rules enable efficient and effective relational learning, and this thesis presents evidence that this is indeed the case. To this end, a system for generating random relational rules is described, and algorithms using these rules are evaluated. These algorithms include direct classification, classification by propositionalisation, clustering, semi-supervised learning and generating random forests. The experimental results show that these algorithms perform competitively with previously published results for the datasets used, while often exhibiting lower runtime than other tested systems. This demonstrates that sufficient information for classification and clustering is retained in the rule generation process and that learning with random rules is efficient. Further applications of random rules are investigated. Propositionalisation allows single-table algorithms for classification and clustering to be applied to the resulting data, reducing the amount of relational processing required. Further results show that techniques for utilising additional unlabeled training data improve accuracy of classification in the semi-supervised setting. The thesis also develops a novel algorithm for building random forests by making efficient use of random rules to generate trees and leaves in parallel.

In this thesis, a mathematical model that describes a “Unique Constraint Domain” is defined. Following, the “Ordered Unique Constraint Domain” is also mathematically defined. With those definitions, a cross-domain ordering is also defined. Then it is shown that relationships between tables in a Relational Database Management System can be defined in other forms than the usual ways, using cross-domain indexes, based in cross-domain ordering. It is shown that all foreign keys in a database can be transformed in indexes with the benefit of speeding data access. It is also shown that this technique is consistent with actual modeling techniques. It is shown how the index structure, with indexes defined as functions, can provide support for relationship roles. In addition, it is also shown how this can provide support for more than two tables in one relationship and for supporting special sorting order. The addition of a mathematical function to a relation that could sort that relation, demonstrating that the closure property of relations are still kept, shows that this mathematical model can be used as extension of the base relational model. Next, it is shown that with this new technique, commercial database engines should not degrade performance because all supporting structures are already present and, in some cases, a better performance might be achieved. Code for a prototype based in a Commercial Database Engine has been added, as an annex, to show how this new technique can be used.

International Telemetering Conference Proceedings / October 23-26, 2000 / Town & Country Hotel and Conference Center, San Diego, California / TechSat21 is sponsored by the Air Force Research Laboratory Space Vehicles Directorate and the Air Force Office of Scientific Research1. Its mission is to control a cluster of satellites that, when combined, create a “virtual satellite” with which to conduct various experiments in sparse aperture sensing and formation flying. Customers of the TechSat21 database include mission planners and system engineers. Mission Planners need information that allows them to make high level planning and scheduling decisions. System Engineers need information to predict satellite sub-system problems and conduct satellite design and performance trade studies. This paper describes those users and the project database.

No description available.

No description available.

No description available.

Approved for public release; distribution in unlimited. / The relational database model has become the most popular and widespread database model. Most current database systems are based upon or related to -he relational model. However, the relational model is beset with significant limitations, pitfalls and deficiencies. The relational model can be substantially improved with graphical interfaces. To this end, the Graphics Language for Accessing Database (GLAD) can provide easy to use and learn graphics interfaces for the relational model. Data structures and algorithms for GLAD will be presented to extend the relational model. / http://archive.org/details/datastructuresal00gren / Captain, United States Marine Corps

Yu Xiaohui. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 98-115). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Related Work --- p.6 / Chapter 2.1 --- Query Optimization --- p.6 / Chapter 2.2 --- Query Rewriting --- p.8 / Chapter 2.2.1 --- Optimizing Multi-Block Queries --- p.8 / Chapter 2.2.2 --- Semantic Query Optimization --- p.13 / Chapter 2.2.3 --- Query Rewriting in Starburst --- p.15 / Chapter 2.3 --- Plan Generation --- p.16 / Chapter 2.3.1 --- Dynamic Programming Approach --- p.16 / Chapter 2.3.2 --- Join Query Processing --- p.17 / Chapter 2.3.3 --- Queries with Aggregates --- p.23 / Chapter 2.4 --- Statistics and Cost Estimation --- p.24 / Chapter 2.5 --- Histogram Techniques --- p.27 / Chapter 2.5.1 --- Definitions --- p.28 / Chapter 2.5.2 --- Trivial Histograms --- p.29 / Chapter 2.5.3 --- Heuristic-based Histograms --- p.29 / Chapter 2.5.4 --- V-Optimal Histograms --- p.32 / Chapter 2.5.5 --- Wavelet-based Histograms --- p.35 / Chapter 2.5.6 --- Multidimensional Histograms --- p.35 / Chapter 2.5.7 --- Global Histograms --- p.37 / Chapter 3 --- New Histogram Techniques --- p.39 / Chapter 3.1 --- Piecewise Linear Histograms --- p.39 / Chapter 3.1.1 --- Construction --- p.41 / Chapter 3.1.2 --- Usage --- p.43 / Chapter 3.1.3 --- Error Measures --- p.43 / Chapter 3.1.4 --- Experiments --- p.45 / Chapter 3.1.5 --- Conclusion --- p.51 / Chapter 3.2 --- A-Optimal Histograms --- p.54 / Chapter 3.2.1 --- A-Optimal(mean) Histograms --- p.56 / Chapter 3.2.2 --- A-Optimal(median) Histograms --- p.58 / Chapter 3.2.3 --- A-Optimal(median-cf) Histograms --- p.59 / Chapter 3.2.4 --- Experiments --- p.60 / Chapter 4 --- Global Histograms --- p.64 / Chapter 4.1 --- Wavelet-based Global Histograms --- p.65 / Chapter 4.1.1 --- Wavelet-based Global Histograms I --- p.66 / Chapter 4.1.2 --- Wavelet-based Global Histograms II --- p.68 / Chapter 4.2 --- Piecewise Linear Global Histograms --- p.70 / Chapter 4.3 --- A-Optimal Global Histograms --- p.72 / Chapter 4.3.1 --- Experiments --- p.74 / Chapter 5 --- Dynamic Maintenance --- p.81 / Chapter 5.1 --- Problem Definition --- p.83 / Chapter 5.2 --- Refining Bucket Coefficients --- p.84 / Chapter 5.3 --- Restructuring --- p.86 / Chapter 5.4 --- Experiments --- p.91 / Chapter 6 --- Conclusions --- p.95 / Bibliography --- p.98

A materialized view is a derived data collecton stored in a database. When the source data for a materialized view is updated, the materialized view also needs to be updated. The process of updating a materialized view in response to changes in the source data is called view maintenance. There are two methods for maintaining a materialized view - recomputation and incremental computation. Recomputation computes the new view instance from scratch using the updated sources data. Incremental computation on the other hand, computes the new view instance by using the update to the source data, the old view instance, and possibly some source data. Incremental computation is widely accepted as a less expensive mathod of maintaining a view when the size of the update to the source data is small in relation to the size of the source data. / thesis (PhD)--University of South Australia, 2000