Ana Maria de Carvalho Moura

A survey on metadata for describing and retrieving Internet resources

Metadata, or “information that makes data useful,” have been considered by the database community basically as data in dictionaries used to control database management systems operations. More recently, metadata have been used to describe digital resources available across networks. This paper presents a survey of the state of the art concerning the use and importance of metadata, focusing on different standards and models found in the literature, describing how they serve as a basis for integrating heterogeneous resources on the Web and for developing more sophisticated search mechanisms.

The Ecobase project: database and web technologies for environmental information systems

A very large number of data sources on environment, energy, and natural resources are available worldwide. Unfortunately, users usually face several problems when they want to search and use relevant information. In the Ecobase project, we address these problems in the context of several environmental applications in Brazil and Europe. We propose a distributed architecture for environmental information systems (EIS) based on the Le Select middleware developed at INRIA. In this paper, we present this architecture and its capabilities, and discuss the lessons learned and open issues.

Integrating Ontologies Based on P2P Mappings

Large organizations usually have difficulties in dealing with the exponential growth of information. Therefore, there is a high demand for innovative solutions to deal with such growth and to integrate such information. This paper proposes a new approach, called Emergent Ontologies (EOs), toward the generation of a single organizational ontology through which it becomes possible to browse all information of an organization. This proposal considers that, typically, an organizations information is distributed in peers, and that in each peer, this information could be represented through a different ontology. As each peer of an organization needs to exchange information, peer-to-peer mappings are created to bridge these ontologies. Based on these mappings, this paper proposes a set of heuristics, which are used to generate the EO. These heuristics have been incorporated into the OntoEmerge system, a prototype developed to facilitate the creation of an initial organization ontology. In order to evaluate such a system and the heuristics behind it, some experiments have been performed. A quantitative and qualitative analysis of these experiments is also presented in this paper. This approach presents encouraging results, and this fact can be considered as a starting point for the creation of organizational ontologies.

Metadata to Support Transformations and Data & Metadata Lineage in a Warehousing Environment

Data warehousing is a collection of concepts and tools which aim at providing and maintaining a set of integrated data (the data warehouse – DW ) for business decision support within an organization. They extract data from different operational data sources, and after some cleansing and transformation procedures data are integrated and loaded into a central repository to enable analysis and mining. Data and metadata lineage are important processes for data analysis. The first allows users to trace warehouse data items back to the original source item from which they were derived and the latter shows which operations have been performed to achieve that target data. This work proposes integrating metadata captured during transformation processes using the CWM metadata standard in order to enable data and metadata lineage. Additionally it presents a tool specially developed for performing this task.

The ROSA project: leveraging e-learning to a semantic layer

Learning Content Management System (LCMS) supports e-learning applications with storage and efficient access for e-Learning Objects (LOs). In order to foster LCMS adoption by the e-learning community, an important issue is to bring course modelling to the realm of the LCMS. In this paper, we describe the Repository of Objects with Semantic Access (ROSA) project. ROSA is a LCMS that provides users with a conceptual map view of courses. We present ROSA data model designed as an extension to RDF data model with ordered collections, relationship properties and cardinality constraints. An algebra is proposed with operations supporting queries over LO metadata and navigation through conceptual maps. Relationship properties are implicitly considered during navigation. User queries can be expressed through a high-level query language, directed by a browsing interface or based on keyword search over LO attribute values. Additionally, a thesaurus is used during query processing extending the domain vocabulary with synonyms, generic/specific and Associated Terms (ATs). Furthermore, ROSA supports the contextualisation of conceptual maps through an extension based on Topic Maps (TM). Finally, the system provides users with a computable semantic layer representation of conceptual maps, which includes course modelling and course material management, leveraging current LCMS.

ROSA - P2P: a peer-to-peer system for Learning Objects integration on the web

ROSA is an e-learning system which enables the management of Learning Objects (LOs). However, since ROSA is a centralized system, it does not provide the integration of LOs created in other institutions. This paper presents the P2P distributed system developed for ROSA, named ROSA - P2P, which provides the physical environment to carry out the integration of these LOs. This environment includes, among other functionalities, particular strategies for: aggregation of peers and grouping of super-peers; connection/disconnection of peers; definition and election of super-peers; balancing and redistribution of peers; routing indices for peers and super-peers communication; and some issues concerning fault tolerance.

SiBBr: Uma Infraestrutura para Coleta, Integração e Análise de Dados sobre a Biodiversidade Brasileira

In this article we describe the Brazilian Biodiversity Information System, which aims to provide an infrastructure for gathering, integrating, and analyzing data produced by various institutions in this area. Both its architecture and the process for harvesting and indexing data on species occurrences and checklists, one of the already implemented components, are described. An implementation of a scalable scientific workflow for species distribution modeling, one of the most used applications for analyzing biodiversity, is presented. Finally, we describe current work on integrating socioeconomic data and on managing ecological data. Resumo. Neste artigo descrevemos o Sistema de Informação sobre a Biodiversidade Brasileira, que tem como objetivo fornecer uma infraestrutura para coleta, integração e análise de dados produzidos e disponibilizados por diversas instituições da área. A arquitetura de software do sistema é descrita, bem como o fluxo de coleta e indexação de dados de ocorrências e listas de espécies, um dos componentes já implementados. É apresentada a implementação de um workflow científico escalável para a modelagem da distribuição de espécies, uma das aplicações mais utilizadas na análise da biodiversidade. Finalmente, são descritos trabalhos em andamento como a integração de dados socioeconômicos e o gerenciamento de dados ecológicos.

An Ontology-Based Reasoning Approach for Document Annotation

Information management has become an important challenge, especially when most of their relevant and strategic documents are available on the Web only for human interpretation. Annotating documents rises as an interesting strategy to diminish the hard task of retrieving important documents from the Web. Annotations consist of associating metadata with text segments of a document, in order to facilitate its retrieval by search engines. Besides improving their performance, annotations enable an optimized indexation of documents. However, due to the huge amount of existing documents, the idea of generating document annotations is not a trivial task. This paper presents a proposal for automatically enriching documents with semantic annotations, where document terms are annotated according to a domain ontology. Currently there already exist some document annotation tools to automate this process. However, the main contribution of this work is focused on the ability of exploring the ontology inference capability and on the meta-annotation concept, which aim at providing users and automatic agents with a more powerful mechanism to retrieve information.

A metadata approach to manage and organize electronic documents and collections on the web

In recent years, the number of information sources offered on the Web has grown tremedously. Support for accessing these information sources has mostly been concentrated on browsing and search tools. Digital libraries and Web directories constitute important initiatives to improve information access, creating and organizing document collections hierarchically, according to different criteria. Search tools, on the other hand, offer a more comprehensive coverage of resources, using robot-based services to collect and index documents, that can be latter accessed using information retrieval techniques. However, technologies applied to search mechanisms on the Web still offer little support to manage document collections, as the association between these documents cannot be explicitly identified, neither by their formats nor by their types. This paper presents a formal structure for organizing and describing collections and their documents on the Web. It is based on a metadata conceptual model which explores relationships between information resources at different levels of granularity. To validate this model, a prototype has been implemented using both a semi-structured and an object-relational database (DB) approach.

A scientific hypothesis conceptual model

In-silico scientific research is a complex task that involves the management of huge volumes of data and metadata produced during the scientific exploration life cycle, from hypothesis formulation up to its final validation. This wealth of data needs to be structured and managed in a way that readily makes sense to scientists, so that relevant knowledge may be extracted to contribute to the scientific investigation process. This paper proposes a scientific hypothesis conceptual model that allows scientists to represent the phenomenon been investigated, the hypotheses formulated in the attempt to explain it, and provides the ability to store results of experiment simulations with their corresponding provenance metadata. The proposed model supports scientific life-cycle: provenance, scientists exchange of information, experiment reproducibility, model steering and results analyses. A cardiovascular numerical simulation illustrates the applicability of the model and an initial implementation using SciDB is discussed.