Dynamic link prediction method of task and user in Mobile Crowd Sensing

Abstract

To address the problem that the existing Mobile Crowd Sensing task allocation methods only considering the perspective of sensing users, ignoring the potential relationship between tasks and users, this paper proposes a task allocation method based on link prediction. First, we analyze the characteristics of sensing users and sensing tasks in Mobile Crowd Sensing, especially the common attribute characteristics of both, and establish the Mobile Crowd Sensing knowledge graph. Then the deep link relationship between sensing users and sensing tasks is mined by the link prediction method based on knowledge graph reasoning. Finally, according to the predicted fit between users and tasks, high-quality sensing users that meet the requirements of the task are selected, to improve the perceptual quality. The experiment results on different datasets show that the method proposed in this paper is high-quality to other comparison methods in sensing data quality and coverage.