Details of Research Outputs

In the era of big data, the development of mobile Internet and the popularization of mobile terminals have formed massive mobile trajectory data. Reasonable usage of the data will greatly improve the service quality and experience of end users. To analyze hidden activity patterns of end-user in the data, big data query is an important operation and how to enhance the query efficiency remains a challenge issue. However, different data analysis approaches have different applications in different fields, and it is necessary to mine hidden data relationships. In addition, query time is one of important factors to evaluate query efficiency, some researches however mainly focus on query result rather than evaluating query efficiency through multiple contrast approaches. To address these issues, an ontology-based modeling and semantic query strategy for mobile trajectory data is investigated in this paper. First, we respectively employ cosine similarity, point-wise mutual information (PMI) and containment probability model to mine association relationship and containment relationship hidden in the data. Subsequently, an ontology-based model is built to visualize end-user's activity through taxonomy and comparison approaches. Finally, four semantic query methods, e.g., basic query, join query, containment query and combination (join containment) query, are defined through SPARQL (SPARQL Protocol and RDF Query Language) to evaluate query time, and the query efficiency achieved by these investigations has been demonstrated through the conducted experiments.