Personalized micro-learning support based on process mining

Micro-Learning is a new leaning paradigm based on microblogging, e-mail or SMS, in which an integral learning resource consists of a series of micro learning units that are dispersed via the services of Internet, and can be used to help users learn at anywhere with a short-term. But the problem is that most of the micro learning units are disorganized. Therefore, it is a critical issue how to organize these micro learning units, in order to make them easier to be used and learned. In this study, we propose an approach based on process mining to organize the learning units according to the situation of users. Firstly, the successful micro-learning processes are extracted from the access logs. And then, the learning process map named navigation map is created based on the domain knowledge and the access sequence of users. Secondly, according to the similarity of access behavior, a reference user group is extracted dynamically for a target user. Based on the dynamic Bayesian network, the navigation map is then used to calculate the posterior probabilities of learning units. Finally, on the basis of posterior probabilities, a learning unit can be recommended for the target user to learn as the next step. The results are provided to the target user gradually, until the target user finish the whole course by the guide of learning process at his/her fragmented time.