Karim Abbasi - Department of Computer Science University of Sistan and Baluchestan Zahedan, Iran
Amin Rahati - Department of Computer Science University of Sistan and Baluchestan Zahedan, Iran
Hossein Banadkuki - Department of Computer Science University of Sistan and Baluchestan Zahedan, Iran

Intelligent tutoring system is a dynamic personalized environment which makes use of e-learning system's benefits. In this paper, a new approach using Bayesian network is proposed to evaluate and analyze the perception level, knowledge level and the skill level of students in an intelligent tutoring system. Further, the levels of perception, knowledge, and student’s skills are clustered into four groups of very low, low, average, and high utilizing k-means clustering algorithm. Hence, the system has the capacity to suggest applicable recommendations to each student. To evaluate the proposed approach, it has been applied on the recorded factual data which are provided by the students during an Artificial Intelligence (AI) course. The obtained results confirm that our approach is achieved a remarkable performance.

Intelligent Tutoring System, E-learning, Bayesian Network, Clustering