Location-to-trajectory matching plays an important role in trajectory data management and analysis. In this paper, we propose and study a novel problem of privacy-preserving spatial keyword location-to-trajectory matching (PPSKLT Matching). Given a set O of locations with keywords, a set T of activity trajectories, a spatial matching threshold $$\theta _{\text{sp}}$$ θ sp , and a keyword matching threshold $$\theta _{\text{kw}}$$ θ kw , the PPSKLT matching finds all location-trajectory pairs from O and T while preserving the users’ privacy. We believe that the PPSKLT matching may benefit many mobile applications such as trajectory activity analysis, event tracking, and so on. The PPSKLT matching is challenging due to three reasons: (1) how to define the spatial keyword similarity measure between locations and trajectories, (2) how to prune the search space effectively, and (3) how to preserve the users’ privacy during query processing. To overcome these challenges and address the problem efficiently, we develop a novel network expansion algorithm (NEA). We define a pair of upper and lower bounds on the spatio-textual similarity to prune the search space. We also define a privacy-preserving mechanism to protect users’ privacy. We conduct extensive experiments on spatio-textual data sets to verify the performance of the developed algorithms.

中文翻译：

---

隐私保护空间关键字位置到轨迹匹配

位置到轨迹匹配在轨迹数据管理和分析中起着重要作用。在本文中，我们提出并研究了一个新的隐私保护空间关键字位置与轨迹匹配（PPSKLT 匹配）问题。给定带有关键字的位置集合 O、活动轨迹集合 T、空间匹配阈值 $$\theta _{\text{sp}}$$ θ sp 和关键字匹配阈值 $$\theta _{\text {kw}}$$ θ kw ，PPSKLT 匹配从 O 和 T 中找到所有位置轨迹对，同时保护用户的隐私。我们相信 PPSKLT 匹配可以使许多移动应用程序受益，例如轨迹活动分析、事件跟踪等。由于三个原因，PPSKLT 匹配具有挑战性：（1）如何定义位置和轨迹之间的空间关键字相似性度量，(2) 如何有效地修剪搜索空间，以及 (3) 如何在查询处理过程中保护用户的隐私。为了克服这些挑战并有效地解决问题，我们开发了一种新颖的网络扩展算法（NEA）。我们在空间文本相似度上定义了一对上限和下限来修剪搜索空间。我们还定义了隐私保护机制来保护用户的隐私。我们对空间文本数据集进行了大量实验，以验证所开发算法的性能。我们还定义了隐私保护机制来保护用户的隐私。我们对空间文本数据集进行了大量实验，以验证所开发算法的性能。我们还定义了隐私保护机制来保护用户的隐私。我们对空间文本数据集进行了大量实验，以验证所开发算法的性能。