Uma Nova Abordagem para Aplicação de Reforço em Sistemas Automáticos e Adaptativos de Detecção de Estilos de Aprendizagem

Abstract

Techniques for Automatic Detection of Learning Styles have been addressed to improve the performance of students who attend Distance Education. The importance of this Automatic Detection lays in the possibility of creating Virtual Learning Environments with automatic adaptation to the students’ profiles,  thus  providing  better  experiences  and  greater  efficiency  in  the  learning process.  In order to evaluate techniques that aim to detect (and adjust) the Learning Styles from students, this work uses a well-known simulator found in the literature. In this system, combinations of Learning Styles are selected and then the chosen combination is evaluated (simulating its performance) according to the student’s actual learning style.  If the performance is unsatisfactory, then a reinforcement is applied in order to guide the system to find the student’s actual Learning Style.  The objective of this work is to improve the reinforcement applied in this simulator. Results show that there are statistically significant differences and a superiority of the proposed method in relation to the prevailing literature approach.