Reúso de conteúdo da Web na recomendação personalizada de objetos de aprendizagem: uma abordagem baseada em um algoritmo genético, tecnologias da Web Semântica e uma ontologia

Abstract

E-learning is an electronic teaching model that can be adapted to the learning styles of every student. In this context, many of the available educational resources are not properly structured in any pattern that facilitates their reuse. In addition, learning objects (LOs) have certain peculiarities and metadata (information that describes them) that make their creation a time-consuming and costly task. To solve this problem, this work develops an approach that uses Wikipedia content to create new LOs and creates an ontology for modeling LOs and students. The system (SCROA) that implements this approach has two types of recommendations, both of which are assisted by inference rules used to suggest all ontology LOs that have some similarity to user search parameters. In the first type of recommendation, LOs that best meet these parameters are recommended. In the second type of recommendation, the user also defines some concepts that the learner is expected to learn, so there is a concept-based Learning Object Recommendation Problem, which aims at recommending LOs that cover all concepts and at the same time meet the profile of the student and other parameters defined by the user. A genetic algorithm (GA) solves this problem. When the LOs suggested by the ontology are not sufficient to cover all the concepts, before the GA is executed, new LOs are created by the reuse of wiki content taking into account their quality. The GA ensures that the recommended LOs cover all concepts and meet the student profile. Given the efficiency of the SCROA approach, this can influence the popularization of Adaptive Hypermedia Systems that allow students of all social classes to have a better performance in the teaching and learning process.