USING PROFILING FOR HELPING THE TEACHING OF PORTUGUESE WITH AN ARTIFICIAL TUTOR

The last generation of artificial tutors brought a lot of innovative aspects that are being used today in a very diverse range of degrees and levels of education. Even at preschool level we can find their presence acting as complementary tools on the education process of students quite young. Some teachers use artificial tutors to reinforce the learning goals and lesson contents that regularly are established for acquiring in conventional classrooms. In some specific cases, these tools have the ability to follow the learning evolution of a student by keeping detailed records about the things students do when are studying some particular topic or lesson or when they are doing some exam or evaluation test. Basically, they have the ability to establish student profiles based on the interaction with the artificial tutor itself and with the use of its educational resources. In this paper, we present and discuss a student profiling system conceived to be integrated in an artificial tutor that was especially designed and developed for helping teaching the Portuguese language to children aged 4 to 5 years. Our basic goal was to provide a set of mechanisms to provide to the artificial tutor the ability to adapt automatically the educational sessions (reading and writing) established in the method of Cartilha Maternal of ‘João de Deus’ to student learning needs, contributing to a more effective and personalized support to classes and individual plans of learning – one specific learning plan for one specific student. To achieve this goal, it was necessary to develop a set of mechanisms specifically oriented to the establishment of usage profiles - behavior patterns - built from the information held by the artificial tutor about the study sessions done by students. The exploitation of this information has to be systematic and involves a careful analysis of the available data through proven data mining processes. Using Markov chains, we were able to create the basic structures for the establishment of usage profiles of students and check their variation according to the different perspectives of analysis and learning objectives established previously by the teacher. This brought to the artificial tutor the ability to adapt its functionalities and services to the requisites of learning of the students.