Toward mining user movement behaviors in indoor environments

In this paper, we explore a new mining paradigm, called User Visited Patterns (abbreviated as UVP), to discover user visited behavior in the mall-like indoor environment. It is a highly challenging issue, in the indoor environment, to retrieve the frequent UVP, especially when the concern of user privacy is highlighted nowadays. The mining of UVP will face the critical challenge from spatial uncertainty. In this paper, the proposed system framework utilizes the probabilistic mining to identify top-k UVP over uncertain dataset collected from the RFID-based sensing result. Moreover, we redesign the indoor symbolic model to enhance the accuracy and efficiency. Our experimental studies show that the proposed system framework can overcome the impact from location uncertainty and efficiently discover high-quality UVP, to provide insightful observation for marketing collaborations.