Wang, Philips Fu LeePhilips Fu LeeWangWang, Z.Kang, P.Wu, Z.Rao, Y.2021-04-142021-04-142016https://repository.sfu.edu.hk/handle/sfu/538With the rapid development of Web 2.0 and interactive technologies, learning resources are proliferating online. Confronting such large volume of educational data, users require effective and efficient methodologies to organize and manage them, which reveals the importance of clustering. In this paper, we first propose a method to estimate the data density, and then apply it to merge learning resources. The proposed algorithm estimates the confidence of any two learning resources to be a pair of neighbors, and conducts clustering by combining the above confidence with the similarities among resources. Experiments are designed to evaluate the performance of our algorithm using the standard clustering datasets. We also demonstrate how to employ the proposed algorithm in educational applications, including e-learner grouping, resource recommendation and usage patterns discovery.enconference proceedings10.1007/978-3-319-32865-2_13