The quality of metadata in a digital library is an important factor in ensuring access for end-users. Several studies have tried to define quality frameworks and assess metadata but there is little user feedback about these in the literature. As collections grow in size maintaining quality through manual methods becomes increasingly difficult for repository managers. This research presents the design and implementation of a web-based metadata analysis tool for digital repositories. The tool is built as an extension to the Greenstone3 digital library software. We present examples of the tool in use on real-world data and provide feedback from repository managers. The evidence from our studies shows that automated quality analysis tools are useful and valued service for digital libraries.

Die Arbeit setzt sich mit dem Einsatz der Component Metadata Infrastructure CMDI im Rahmen der föderierten Infrastruktur CLARIN auseinander, wobei diverse konkrete Problemfälle aufgezeigt werden. Für die Erarbeitung entsprechender Lösungsstrategien werden unterschiedliche Verfahren adaptiert und für die Qualitätsanalyse von Metadaten und zur Optimierung ihres Einsatzes in einer föderierten Umgebung genutzt. Konkret betrifft dies vor allem die Übernahme von Modellierungsstrategien der Linked Data Community, die Übernahme von Prinzipien und Qualitätsmetriken der objektorientierten Programmierung für CMD-Metadatenkomponenten, sowie den Einsatz von Zentralitätsmaßen der Graph- bzw. Netzwerkanalyse für die Bewertung des Zusammenhalts des gesamten Metadatenverbundes. Dabei wird im Rahmen der Arbeit die Analyse verwendeter Schema- bzw. Schemabestandteile sowie die Betrachtung verwendeter Individuenvokabulare im Zusammenspiel aller beteiligten Zentren in den Vordergrund gestellt.

Cultural heritage institutions, including galleries, libraries, museums, and archives are increasingly digitizing physical items and collecting born-digital items and making these resources available on the Web. Metadata plays a vital role in the discovery and management of these collections. Existing frameworks to identify and address deficiencies in metadata rely heavily on count and data-value based metrics that are calculated over aggregations of descriptive metadata. There has been little research into the use of traditional network analysis to investigate the connections between metadata records based on shared data values in metadata fields such as subject or creator. This study introduces metadata record graphs as a mechanism to generate network-based statistics to support analysis of metadata. These graphs are constructed with the metadata records as the nodes and shared metadata field values as the edges in the network. By analyzing metadata record graphs with algorithms and tools common to the field of network analysis, metadata managers can develop a new understanding of their metadata that is often impossible to generate from count and data-value based statistics alone. This study tested application of metadata record graphs to analysis of metadata collections of increasing size, complexity, and interconnectedness in a series of three related stages. The findings of this research indicate effectiveness of this new method, identify network algorithms that are useful for analyzing descriptive metadata and suggest methods and practices for future implementations of this technique.

Data obtained from foreign data sources often come with only superficial structural information, such as relation names and attribute names. Other types of metadata that are important for effective integration and meaningful querying of such data sets are missing. In particular, relationships among attributes, such as foreign keys, are crucial metadata for understanding the structure of an unknown database. The discovery of such relationships is difficult, because in principle for each pair of attributes in the database each pair of data values must be compared. A precondition for a foreign key is an inclusion dependency (IND) between the key and the foreign key attributes. We present with Spider an algorithm that efficiently finds all INDs in a given relational database. It leverages the sorting facilities of DBMS but performs the actual comparisons outside of the database to save computation. Spider analyzes very large databases up to an order of magnitude faster than previous approaches. We also evaluate in detail the effectiveness of several heuristics to reduce the number of necessary comparisons. Furthermore, we generalize Spider to find composite INDs covering multiple attributes, and partial INDs, which are true INDs for all but a certain number of values. This last type is particularly relevant when integrating dirty data as is often the case in the life sciences domain - our driving motivation.

Metadata is essential to how (digital) archives, collections or databases operate. It is the backbone to organise different types of content, make them discoverable and keep the digital records’ authenticity, integrity and meaning over time. For that reason, it is also important to iteratively assess if the metadata is of high quality. Despite its importance, there is an acknowledged lack of research verifying if existing assessment frameworks and methodologies do indeed work and if so how well, especially in fields outside the libraries. Thus, this thesis conducted an exploratory case study and applied already existing frameworks in a new context by evaluating the metadata quality of audiobooks. The Information Continuum Model was used as a way to capture the metadata quality needs of customers/end users who will be searching and listening to audiobooks. Using a mixed methods approach, the results showed that the frameworks can indeed be generalised and adapted to a new context. However, although the frameworks helped measure, identify and find potential solutions to the problems, they could be better adjusted to the context and more metrics and information could be added. Thus, there can be a generalised method to assess metadata quality. But the method needs improvements and to be used by people who understand the data and the processes to reach its full potential.

Submitted by Irene Nascimento (irene.kessia@ufpe.br) on 2016-10-17T18:07:42Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Dissertação - Everaldo Costa Silva Neto (final).pdf: 1973752 bytes, checksum: 18ff29972829bab54f92cc990addf923 (MD5) / Made available in DSpace on 2016-10-17T18:07:42Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Dissertação - Everaldo Costa Silva Neto (final).pdf: 1973752 bytes, checksum: 18ff29972829bab54f92cc990addf923 (MD5) Previous issue date: 2016-08-24 / Atualmente, um massivo volume de dados tem sido produzido pelos mais variados tipos de fontes de dados. Apesar da crescente facilidade de acesso a esses dados, identificar quais fontes de dados são mais adequadas para um determinado uso é um grande desafio. Isso ocorre devido ao grande número de fontes de dados disponíveis e, principalmente, devido à ausência de informações sobre a qualidade dos dados. Nesse contexto, a literatura oferece diversos trabalhos que abordam o uso de critérios de Qualidade da Informação (QI) para avaliar fontes de dados e solucionar esse desafio. No entanto, poucos trabalhos consideram o aspecto dinâmico das fontes na etapa da avaliação da qualidade. Nesta dissertação, abordamos o problema de avaliação da qualidade em fontes de dados dinâmicas, ou seja, fontes de dados cujo conteúdo pode sofrer modificações com alta frequência. Como contribuição, propomos uma estratégia onde os critérios de QI são avaliados de forma contínua, com o objetivo de acompanhar a evolução das fontes de dados ao longo do tempo. Além disso, propomos a criação de um Perfil de Qualidade, que consiste de um conjunto de metadados sobre a qualidade de uma fonte, onde seu uso pode ser aplicado para diversos fins, inclusive no processo de seleção de fontes de dados. O Perfil de Qualidade proposto é atualizado periodicamente de acordo com os resultados obtidos pela avaliação contínua da qualidade. Dessa forma, é possível refletir o aspecto dinâmico das fontes. Para avaliar os resultados deste trabalho, mais especificamente a estratégia de avaliação contínua da qualidade, utilizamos fontes de dados do domínio Meteorológico. Os experimentos realizados demonstraram que a estratégia de avaliação proposta produz resultados satisfatórios. / Nowadays, a massive data volume has been produced by a variety of data sources. The easy access to these data presents new opportunities. In this sense, choosing the most suitable data sources for a specific use has become a challenge. Several works in the literature use Information Quality as a mean of solving this problem, however, only few works employ a continuous strategy. In this work, we address the problem of performing assessment continuously, looking to dynamic data sources. We also propose the creation of a data source Quality Profile, which consists of a set of metadata about the data source’s quality and may be used to help the selection of data sources. To reflect the real quality values of a data source, we propose a continuous updating of the Quality Profile, according to the data source’s refresh rate. In order to evaluate our proposal, we carried out some experiments with meteorological data provided by institutions that monitor weather conditions of Recife. The experimental results have demonstrated that our strategy produces more satisfactory results than others, regarding the trade off between performance and accuracy.

Contrairement à l'opinion dominante, les nouvelles technologies n'ont pas toujours un impact positif sur la qualité des métadonnées dans le secteur culturel. Après dix ans d'expérience avec les projets de numérisation dans nos musées, bibliothèques et archives, une réflexion critique se montre plus que jamais nécessaire pour évaluer à quelles conditions ce genre de projets d'informatisation peuvent offrir une valeur ajoutée pour la documentation de notre patrimoine culturel. Cette réflexion se base, entre autres, sur un ensemble de case studies représentatifs dans un contexte international. A cette fin, nous présenterons et définirons un cadre méthodologique et conceptuel original concernant l'impact des technologies sur la qualité des métadonnées. Sur cette base, nous proposons et expérimentons trois approches opérationnelles novatrices en vue d'améliorer la qualité des systèmes d'information déployés dans le secteur culturel. / Doctorat en Information et communication / info:eu-repo/semantics/nonPublished

Die Arbeit setzt sich mit dem Einsatz der Component Metadata Infrastructure CMDI im Rahmen der föderierten Infrastruktur CLARIN auseinander, wobei diverse konkrete Problemfälle aufgezeigt werden. Für die Erarbeitung entsprechender Lösungsstrategien werden unterschiedliche Verfahren adaptiert und für die Qualitätsanalyse von Metadaten und zur Optimierung ihres Einsatzes in einer föderierten Umgebung genutzt. Konkret betrifft dies vor allem die Übernahme von Modellierungsstrategien der Linked Data Community, die Übernahme von Prinzipien und Qualitätsmetriken der objektorientierten Programmierung für CMD-Metadatenkomponenten, sowie den Einsatz von Zentralitätsmaßen der Graph- bzw. Netzwerkanalyse für die Bewertung des Zusammenhalts des gesamten Metadatenverbundes. Dabei wird im Rahmen der Arbeit die Analyse verwendeter Schema- bzw. Schemabestandteile sowie die Betrachtung verwendeter Individuenvokabulare im Zusammenspiel aller beteiligten Zentren in den Vordergrund gestellt.