A model for motivation assessment in intelligent tutoring systems

Motivation has an undeniable role in the effectiveness of intelligent tutoring systems. In this research, a model is proposed to integrate students' motivation in intelligent tutoring systems. This model is based on the ARCS Model of Motivational Design and log file analysis to estimate students' motivation. Through expert analysis, it was determined that seven attributes (task time, grade, task difficulty, student's interest in the subject, accordance between content presentation and student's learning style, student's skill level and previous motivational state) affect motivation directly and must be included in the model. In order to determine how accurately these attributes can assess the motivational state of students, a reading comprehension test environment was created using Moodle. Fourteen users participated in the study. Random Forest algorithm was used to classify the collected data into “motivated” and “unmotivated” classes. The correct classification rate was 61%. Although the data set is not big enough, however, this preliminary result show that the model is promising and can be further tested and improved.