This paper presents a model for organization of educational content in connectionists intelligent tutoring systems. The availability of educational content in a single format has emerged as a problem for many students. The inadequacy of a single format of content that ignores differences in individual profiles may have inefficient outcomes in the teaching-learning process. The proposed multilevel structure allows for different combinations of concepts for presentation to the same content. Assuming that the pattern of successful organization of the study by a student can be applied to other students with similar profiles, a system was structured to assist in the task of customization reactive content. The customization is provided by a neural network that links the student's profile to a proximal learning pattern. This pattern is combined with expert rules to enable a probabilistic selection so that the system presents the reactivity in the different learning stages. The results of experiments indicate that the approach is effective in providing better use of the content in the personal study and its potential use in Distance Education.
Didactics personalized contents; learning proximity patterns; intelligent tutor systems; artificial neural networks; didactic personalization
