A ranking fusion approach for RDF keyword-based search systems

Nowadays most people use keyword search to retrieve information and surf the web wherever they are. Ease of use given, this technology has been introduced, in the last decade, also for searching structured data such as RDF graphs, which non-expert users accustomed to web search cannot easily get access to. Indeed, users would need to use tools like SPARQL that require a knowledge of the structure of the data it is being searched. In order to apply keyword search over structured data, virtual documents approaches and text retrieval technologies have been proposed. In particular, in this work, we refer to a previous study on the topological semantical aggregator and virtual document pruning over RDF datasets: these two technologies produce state-of-the-art results in terms of efficiency and effectiveness and ensure scalability over millions of triples. This work expands the previous study on RDF keyword search systems modifying and developing new systems for aggregating documents during the offline phase. The additional rankings produced by these systems are then merged following a round-robin rule.