Extracting URI Patterns from SPARQL Endpoints
Understanding the structure of identifiers in a particular dataset is critical for users/applications that want to use such a dataset, and connect to it. This is especially true in Linked Data where, while benefiting from having the structure of URIs, identifiers are also designed according to specific conventions, which are rarely made explicit and documented. In this paper, we present an automatic method to extract such URI patterns which is based on adapting formal concept analysis techniques to the mining of string patterns. The result is a tool that can generate, in a few minutes, the documentation of the URI patterns employed in a SPARQL endpoint by the instances of each class in the corresponding datasets. We evaluate the approach through demonstrating its performance and efficiency on several endpoints of various origins.