Extraction of relationships between page transition and understanding for contents

In this paper, we consider a learning environment, which students solve the problem while reading digital contents. Here, we extract the relationship between the frequency for reading each page and the understanding for given problems. The above relationship can be modeled by using neural networks. Furthermore, we use a learning method with forgetting for neural networks for the purpose of clarification of internal representations. Finally, we have analyzed the relationship between the frequency for each page and the understanding for given contents based on the internal representations of multi-layered neural networks.