Quality Awareness over Graph Pattern Queries

We examine the problem of quality awareness when querying graph databases. According to quality annotations that denote quality problems appearing in data subgraphs (the annotations typically result from collaborative practices in the context of open data usage like e.g. users' feedbacks), we propose a notion of quality aware (graph pattern) query based on (usage-dependent) quality profiles. In this paper, we present the formal foundations of the approach. We also show how to simply extend a generic state-of-the-art algorithm for graph pattern queries evaluation in order to implement quality awareness at evaluation time and we study its complexity. We then expose implementation guidelines, supported by a proof-of-concept prototype based on the Neo4J graph database management system.