Recomendação de objetos de aprendizagem de línguas baseada em inteligência de enxames

Abstract

This work proposes the application of swarm intelligence to recommend language learning objects. To the best of our knowledge, one cannot find in the literature specific works addressing the recommendation of this kind of object. Existing approaches are focused to recommend general learning objects without considering the aspects that guide the learning of languages. In this sense, a study of the essential elements that define a language learning object and the main approaches used in the mentioned general-use recommender systems was conducted. The study considered the application of our approach in a repository developed based on communicative teaching principles and took into account the complex nature of language learning. The recommendation in this context considers the existence of multiple conceptual connections among objects, which are built as the system is used, and are driven by the application of an algorithm based on Ant Colony Optimization. In order to validate our approach, we also implemented two of the general recommendation algorithms most widely used currently. Whereas a language learning object aims to be a facilitator of grammatical skills integration process, sociolinguistic and strategic, two essential aspects must be considered during the recommendation process: the ability to combine an object with another and its reusability, more specifically, the possibility of its application in different learning contexts. Thus, the results obtained from the experiments demonstrated that the proposed approach is efficient for the application context that it is proposed.