Comparison Queries for Uncertain Graphs

Extending graph models to incorporate uncertainty is important for many applications, including disease transmission networks, where edges may have a disease transmission probability associated with them, and social networks, where nodes may have an existence probability associated with them. Analysts need tools that support analysis and comparision of these uncertain graphs. To this end, we have developed a prototype SQL-like graph query language with emphasis on operators for uncertain graph comparison. In order to facilitate adding new operators and to enable developers to use existing operators as building blocks for more complex ones, we have implemented a query engine with an extensible system architecture. The utility of our query language and operators in analyzing uncertain graph data is illustrated using two real world data sets: a dolphin observation network and a citation network. Our approach serves as an example for developing simple query languages that enables users to write their own ad-hoc uncertain graph comparison queries without extensive programming knowledge.