Discovery of action patterns and user correlations in task-oriented processesfor goal-driven learning recommendation

With the high development of social networks, collaborations in a socialized web-based learning environment has become increasing important, which means people can learn through interactions and collaborations in communities across social networks. In this study, in order to support the enhanced collaborative learning, two important factors, user behavior patterns and user correlations, are taken into account to facilitate the information and knowledge sharing in a task-oriented learning process. Following a hierarchical graph model for enhanced collaborative learning within a task-oriented learning process, which describes relations of learning actions, activities, sub-tasks and tasks in communities, the learning action pattern and Goal-driven Learning Group, as well as their formal definitions and related algorithms, are introduced to extract and analyze users' learning behaviors in both personal and cooperative ways. In addition, a User Networking Model, which is used to represent the dynamical user relationships, is proposed to calculate user correlations in accordance with their interactions in a social community. Based on these, an integrated mechanism is developed to utilize both user behavior patterns and user correlations for the recommendation of individualized learning actions. The system architecture is described finally, and the experiment results are presented and discussed to demonstrate the practicability and usefulness of our methods.