Neural Techniques to Improve the Formative Evaluation Procedure in Intelligent Tutoring Systems

Nowadays a special attention is paid to the quality of teaching valued in terms of efficacy of students' knowledge at the exit of its formative period. This efficacy depends both on the quality of formative iter and on the way formative activities are planned in a single teaching. Not since a lot of time has re-emerged the importance of tutoring as ad hoc regulator in student's learning process. Taking care of a student means design both of diagnostic instruments and of suitable formative interventions. That is the moment in which having the diagnosis to be preceded by a survey of performance indicators becomes important the study of measuring system models to be applied to each student. Intelligent tutoring systems (ITS), are any computer system with the ability to adopt pedagogical activities to individual student needs providing customized instruction and feedback. A correct and effective use of such system is obtained through a careful planning of formative deficit diagnostic measuring. In this paper we propose a model of formative evaluation procedure based on the use of neural techniques qualified to improve its accuracy.