Queries with aggregate functions over fuzzy RDF data

Since the World Wide Web Consortium (W3C) recommended the Resource Description Framework (RDF) data model, it has been widely adopted and applied in the era of Big Data and knowledge engineering. Now, large volumes of RDF data are produced and made available regularly. In real-world applications, information is often vague or ambiguous. Therefore, data may contain imperfect information, particularly in the context of open information extraction from the Web. The fuzzy RDF model has been proposed to handle uncertain semantics in RDF data. Aggregate functions are a group of functions for statistics and group data calculation, which are essential for the analysis and processing of RDF data. In spite of the increased attention given to aggregate queries over RDF data, little has been done on aggregate queries over fuzzy RDF data. This paper proposes a more scalable and flexible aggregate query method for fuzzy RDF data. By introducing the concepts of upper and lower bounds and expectation in probability theory, we identify three variants of the aggregate functions in RDF query language over the fuzzy RDF model. The language has been implemented by transformation based on standard SPARQL language and evaluated with experiments. In addition, we examine how to handle multiple types of aggregate operations.