Search is this category of applications which is addressed by vendors of semantic technologies most frequently. Since 'Semantic search' serves only as a general paraphrase for search beyond full-text search, the offered functionalities, used technologies, and resulting advantages can vary widely.
While most methods of text mining have traditionally facilitated semantic search for documents, linked data protagonists also propagate the paradigm shift towards a 'semantic search for knowledge'.
Is this kind of semantic search which goes beyond 'searching for documents' the only 'true' semantic search? What are the use case scenarios, which benefit most from search applications based on linked data and knowledge graphs?
Following the 5-star rating system for linked open data by Tim Berners-Lee, we will present a model which helps to classify the various forms of semantic search. For each of the categories we will give concrete examples and we will discuss technological approaches.
Starting with some advanced features to improve search for documents like auto-complete or query expansion, we will continue with the examination of various forms of search assistants (like search facets) and contextualization based on knowledge graphs. Finally we will explain why the highest rating is given only to search systems based on SPARQL.
Attendees of this talk will receive a comprehensive overview over the state-of-the-art of semantic search technologies and they will learn how to classify them. Upcoming trends in this sector like the usage of knowledge graphs, linked data and SPARQL will be illustrated by specific examples.