APPLICATION OF RECOMMENDER SYSTEMS ON E-LEARNING ENVIRONMENT

In recent years, many advances in educational systems have occurred in order to introduce new technologies such as web based training. Nowadays, many people have benefited from various elearning applications. However, high diversity of the learners on the Internet poses new challenges to the traditional " one- size- fit- all" learning model, in which a single set of learning resources is provided to all learners. In fact, the learners may have different levels of expertise, and hence they cannot be treated in a uniform way. Recommender Systems are used to avoid this problem, increase the efficiency of e- learning environment and enhance the quality of education and motivation of learners. An e- learning recommender system would recommend a learning task to a student based on his behavioral pattern, and based on tasks performed by other similar students. The similarity of the students could be established using user profiles, or could be based on common previous access patterns. In principle, there are two major parts in the design of such a system: a " learning" module that learns from past access patterns and infers an individual or common access model; and an " advising" module that applies the learned model at given times to recommend actions. In this research we have developed a recommender system capable of providing students with appropriate educational materials that suit their different levels, therefore prevent at- risk students from failing and improves their academic achievement. We are interested in recommending beneficial learning activities to enhance online learning, as well as recommending shortcuts or jumps to some resources to help users better navigate the course materials. The system uses data mining techniques in order to analyzing students' reading data and predicting their final results based on the similar students' records before completing their course. For this purpose the WEKA toolkit is used. In this step we build a classification model using the decision tree method. After the process of recognizing weak and strong students, different approaches are then used to provide recommendations. If the system detects that the new student is strong and capable of successfully completing the course, it uses clustering approach. In each cluster, we assume that the students with greater knowledge (for example, those who have obtained better results in various tests) have greater influence in providing of recommendations. If the system finds out that the new student is a weak student or is predicted to fail, it applies the other approach (association rule mining), which is particularly used for weak students. In order to evaluate the system's performance, probability of correct prediction of students' results is calculated. The experimental results indicate that the proposed system is effective in terms of making appropriate recommendations for weak students in order to increase their success rate, as well as for strong students in order to enhance their performance. For example, those weak students who are predicted to fail are able to pass the final test, with receiving the recommendations before completing their course. On the other hand, the probability of successfully passing the course by strong and good students has also increased. The experimental results show that such students show a better performance under our recommendation system.