Personalized Learning Path Generation based on Network Embedding and Learning Effects

With the development of network technology and data mining technology, personalized learning path recommendation in intelligent teaching system has become an important issue. Choosing the right learning object and recommending it to learners is a challenge for online learning systems. At present, the problem faced by online learning is that the information overload causes the learner cannot find a learning resource suitable for himself, and the recommended learning resource cannot satisfy the individual needs of the learner. The recommendation system is an important means of information filtering and a very promising method for solving information overload problems. Therefore, adaptive learning path recommendation is considered to be an effective means to solve the above problems in online learning. In this paper, we propose a learning path generation algorithm based on network embedding and learning effects. First, we use the network embedding to learn the user's expression in order to judge the learner's similarity. Then we generate an adaptive learning path for current learners based on the similarity between all historical learners and current learners and the learning effects of historical learners. The experimental results show that the average grade of the learners using our recommended learning path has been greatly improved.