A Similarity Search of Trajectory Data Using Textual Information Retrieval Techniques

In this paper, we propose a novel similarity measure between trajectory data and geometry data using textual information retrieval techniques. Currently, many trajectory data are generated and used for sightseeing. When users search trajectory data at sightseeing, if a user's current position is similar to a retrieval target trajectory datum, this trajectory datum should be useful. However, even if the euclidean distances between the moving points in trajectory data and the user's position have small values, these trajectory data are not always relevant to the user's interests. In this paper, we deal with textual information retrieval method to measure the similarity values between retrieval target trajectory data and user's current position. Using our proposed method, users can gain relevant sightseeing spots as retrieval results. In our experimental evaluation, we confirmed that our proposed method can retrieve intuitively relevant trajectory data.