III-SGER: Spatio-Temporal-Thematic Queries of Semantic Web Data: a Study of Expressivity and Efficiency

空间和时间数据是许多应用程序中的关键组成部分。在从科学发现到国家安全和刑事调查的分析应用中尤其如此。这项探索性研究开发了空间、时间和主题（STT）数据建模和查询的新方法。这些方法与传统的STT数据管理方法有很大不同；它们遵循的范式超越了查询资源，而是查询资源之间的关系。这将导致三个STT数据管理方面的进步：(1)利用语义Web数据模型以图为中心的特性的新查询操作符；(2)专门用于语义Web数据模型的STT数据的新索引和查询处理技术；(3)扩展SPARQL RDF查询语言以支持STT查询。这个项目的第二个方面是将上面描述的STT-RDF方法与基于OWL-DL和定性空间和时间推理的替代方法进行比较。本探索性研究评估了STT-RDF分析方法是否比具有时空本体的OWL-DL方法提供了一种更高效、更富表现力的查询语言。具体来说，(1)对于可以在两种形式中编码的查询，前一种实现是否更有效？(2)是否存在可以在STT-RDF分析形式化中表述而不能表示为OWL-DL查询的查询？在整个项目中，特别注意实验结果的可重复性。所有代码都将是开源的，所有基准测试、数据集和本体都将通过项目网站athttp://knoesis.wright.edu/research/semweb/projects/stt/提供

Spatial and temporal data are critical components in many applications. This is especially true in analytical applications ranging from scientific discovery to national security and criminal investigation. This exploratory research develops of new methods for modeling and querying spatial, temporal and thematic (STT) data. The methods differ significantly from traditionalapproaches for STT data management; they follow a paradigm that goes beyond querying for resources to querying about the relationships between resources.Three STT data management advances this will lead to are: (1) new query operators that exploit the graph-centric nature of Semantic Web data models, (2) new indexing and query processing techniques for STT data that are specialized for Semantic Web data models and (3) an extension of the SPARQL RDF query language to support STT queries.A second aspect of this project is to compare the STT-RDF approach described above with an alternative approach based on OWL-DL and qualitative spatial and temporal reasoning. This exploratory study evaluates whether the STT-RDF analytics approach provides a more efficient and expressive query language than the OWL-DL approach with space-time ontology. Specifically, (1) for the queries that can be encoded in both the formalisms, is the former implementation more efficient? and (2) are there queries that can be formulated in the STT-RDF analytics formalismthat cannot be expressed as OWL-DL queries? Throughout this project, special attention is given torepeatability of experimental results. All code will be open source, and all benchmarks, datasets and ontologies will be available through the project web site athttp://knoesis.wright.edu/research/semweb/projects/stt/