SEQUENCIAMENTO E RECOMENDAÇÃO DE AÇÕES PEDAGÓGICAS BASEADOS NA TAXONOMIA DE BLOOM E NO PERFIL RASI USANDO PLANEJAMENTO AUTOMATIZADO POR ALGORITMO GENÉTICO Genético

Abstract

The sequencing and recommendation of personalized pedagogical actions in virtual learning environments are two relevant aspects in the attempt to promote and make effective computer-mediated teaching. Thus, this work investigates the use of Artificial Intelligence (AI) planning for the sequencing of these actions, according to the student's profile. The actions are modeled to correspond to the cognitive process described by Bloom's Taxonomy, and the student's profile is determined according to the Revised Approaches to Studying Inventory (RASI). The basic principles that guide Bloom's Taxonomy and the RASI are equivalent; however, it was necessary to map these two theories to measure the adherence of a sequence of actions to the student's profile, and this mapping is one of the contributions of this work. Since planning is a task with a high degree of complexity, the use of evolutionary computing techniques, such as genetic algorithms, and the formulation of the problem as an optimization problem can help in the search for good solutions (sequences of pedagogical actions), as demonstrated in this work. To this end, it was necessary to propose two objective functions in a multi-objective genetic algorithm to evaluate the sequence during the evolution of the algorithm. The recommended actions are those from Bloom's Digital Taxonomy, according to their relevance in each state of the cognitive process. Experiments carried out included higher education students who answered the RASI questionnaire and, after receiving the respective sequences of actions determined by the planner proposed in this work, also answered a satisfaction questionnaire regarding the resulting sequence of pedagogical actions. Such results were promising and point to the viability of the proposal, with potential to compose virtual learning environments.