Full text
| |
| Author(s): |
Luiz Celso Gomes Junior
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Computação |
| Defense date: | 2015-10-26 |
| Examining board members: |
André Santanchè;
Carmem Satie Hara;
Ricardo da Silva Torres;
Lucas Francisco Wanner;
João Eduardo Ferreira
|
| Advisor: | André Santanchè |
| Abstract | |
Understanding and quantifying the emergent properties of natural and man-made networks such as food webs, social interactions, and transportation infrastructures is a challenging task. The complex networks field was developed to encompass measurements, algorithms, and techniques to tackle such topics. Although complex networks research has been successfully applied to several areas of human activity, there is still a lack of common infrastructures for routine tasks, especially those related to data management. On the other hand, the databases field has focused on mastering data management issues since its beginnings, several decades ago. Database systems, however, offer limited network analysis capabilities. To enable a better support for complex network analysis tasks, a database system must offer adequate querying and data management capabilities. This thesis advocates for a tighter integration between the areas and presents our efforts towards this goal. Here we describe the Complex Data Management System (CDMS), which enables explorative querying of complex networks through a declarative query language. Query results are ranked based on network measurements assessed at query time. To support query processing, we introduce the Beta-algebra, which offers an operator capable of representing diverse measurements typical of complex network analysis. The algebra offers opportunities for transparent query optimization through query rewritings, proposed and discussed here. We also introduce the mapper mechanism for relationship management, which is integrated in the query language. The flexible query language and data management mechanisms are useful in scenarios other than complex network analysis. We demonstrate the use of the CDMS in applications such as institutional data integration, information retrieval, classification and recommendation. All aspects of the proposal are implemented and have been tested with real and synthetic data (AU) | |
| FAPESP's process: | 12/15988-9 - Querying and Managing Complex Data |
| Grantee: | Luiz Celso Gomes Junior |
| Support Opportunities: | Scholarships in Brazil - Doctorate |