Enriquecimiento semántico de metadatos educativos apoyado en ontologías de conocimiento general

• Autores: M. Elena Rodríguez
• Directores de la Tesis: Miguel Ángel Sicilia Urbán (dir. tes.) , Jordi Conesa (codir. tes.)
• Lectura: En la Universidad de Alcalá ( España ) en 2012
• Idioma: español
• Tribunal Calificador de la Tesis: Angélica de Antonio Jiménez (presid.) , Salvador Sánchez Alonso (secret.) , Antoni Urpí Tubella (voc.) , María Antonia Huertas Sánchez (voc.) , Manuel Prieto (voc.)
• Enlaces
  • Tesis en acceso abierto en: e_Buah
• Resumen
  • español

    La creación y compartición de contenidos en e-learning constituye una cuestión de interés central. La creación se puede realizar desde cero o partir de la reutilización de contenidos que quizá fueron concebidos para un contexto de uso educativo diferente al de donde se pretenden reutilizar. Dichos contenidos educativos reciben el nombre de objetos de aprendizaje. Para posibilitar la reutilización de los objetos de aprendizaje, es necesario proporcionar mecanismos que faciliten su búsqueda, selección y examen. Estos mecanismos se basan en el uso de los metadatos propuestos por diferentes organismos, siendo el estándar de referencia IEEE LOM.

    LOM incluye un modelo de datos que define de manera descriptiva el conjunto de metadatos que describen los objetos de aprendizaje y proporciona las correspondencias que permiten expresar dicho modelo en XML. La definición descriptiva de LOM causa que esté sujeto a múltiples interpretaciones, dado que contiene imprecisiones, omisiones e información que está oculta y diseminada en diferentes metadatos. Todo ello limita la gestión de los objetos de aprendizaje, e introduce la necesidad de una representación formal que permita que el conocimiento relativo a los objetos de aprendizaje pueda ser definido, interpretado, compartido e intercambiado de manera no ambigua.

    En el trabajo se realiza la definición, mediante ontologías, de un modelo de conocimiento del dominio de los objetos de aprendizaje que permite mejorar la expresividad semántica de LOM y que solventa sus principales ambigüedades y deficiencias. Todo ello ayuda a mejorar las capacidades de los procesos relativos a la gestión de los objetos de aprendizaje. Entre estos procesos se encuentran, además de los antes mencionados, la anotación de los objetos de aprendizaje y la realización de tareas de control y auditoría. El modelo de conocimiento se apoya parcialmente en el uso de ontologías de propósito general e incluye como contribuciones principales, por un lado, un marco conceptual independiente de las tecnologías que permitirían su implementación, y por otro, el desarrollo de un esquema formal basado en ontologías para la representación del conocimiento sobre los objetos de aprendizaje ajustado al marco conceptual.

    La evaluación del esquema formal desarrollado se ha realizado desde diferentes ámbitos. Primero se ha evaluado que el esquema formal está bien descrito y no contiene inconsistencias. En segundo lugar, se ha evaluado su completitud a través del diseño de diversos casos de ejemplo y de su comparación con otros esquemas formales disponibles. Finalmente, en tercer lugar, se ha evaluado su utilidad desde un punto de vista funcional, a través de diversos escenarios de aplicación.

  • English

    The use of Information and Communication Technologies (ICT) in education has enabled the development of a wide range of solutions designed to facilitate teachinglearning processes. Amongst them, there is the feasibility of offering distance and blended learning in the form of e-learning.

    A basic element in e-learning is the creation and sharing of educational content stored in repositories. The creation of educational content can be done from scratch or by total or partial reuse of previously developed educational content. This content, referred to as learning objects, perhaps was conceived for an educational context different to the one for which it is intended to be reused. In order to enable the reuse of learning objects, it is necessary to provide mechanisms that facilitate their search, selection and browsing. These mechanisms rely on the use of metadata that describe the learning objects. Such metadata are proposed by different standards and specifications, the most prominent one being the IEEE 1484.12.1-2002 Standard for Information Technology–Education and Training Systems– Learning Objects and Metadata, also known as the LOM standard.

    The LOM standard includes a data model that defines in a descriptive manner the set of metadata to be used in the characterization of the learning objects. In addition, it also provides (in a separated document) the bindings that specify how the data model can be expressed in XML. For a given learning object, the chosen metadata and their values, together with their structure according to LOM, generate a metadata record that describes the learning object. The descriptive definition of the standard could cause multiple interpretations, since it contains omissions, lack of precision, and information that remains hidden and scattered in different metadata. These shortcomings reduce the capabilities of several processes related to the management of learning objects (e.g. search, selection and browsing), even though the metadata records that describe them may be correctly structured from a syntactic point of view. Such problems introduce the need for a formal representation that allows all the knowledge associated with the learning objects to be defined, interpreted, shared and exchanged in a non ambiguous way. Since e-learning can be considered a particular application domain of the semantic Web that is focused on teaching-learning environments, the development of ontologies (the essential element to provide meaning in the context of the semantic Web) seems to be an appropriate solution.

    This dissertation applies software engineering and artificial intelligence techniques to define a knowledge model of the learning objects domain that improves the semantic expressiveness of the representation schema originally suggested by the LOM standard. This knowledge model conforms to LOM and resolves the main ambiguities and deficiencies that exist in the definitions provided by the standard. Furthermore, the developed knowledge model is designed to improve the management of the learning objects available in repositories. In addition to processes previously mentioned, this knowledge model ensures the consistent annotation (or description) of the learning objects, as well as the accomplishment of control and audit tasks on them.

    The knowledge model leans partially on the use of commonsense ontologies resulting in two significant achievements. Firstly, it contributes to the definition of a conceptual framework independent of the technologies used for its implementation, ensuring the conclusions reached remain intact and valid. Secondly, it contributes to the development of a formal schema based on ontologies for the representation of the knowledge related to learning objects. This formal schema is adjusted to the conceptual framework.

    The formal schema has been evaluated from different perspectives. To begin with, its correctness has been evaluated from a logical-formal point of view, with the aim of ensuring that the formal schema is well described and does not contain inconsistencies. Following this, an evaluation of its completeness is provided by designing different case studies and through its comparison with other available formal schemas. Finally, its usefulness from a functional point of view has been evaluated, suggesting different scenarios where it could be applied.