Special issue on new trends for ontology‐based knowledge discovery

In recent Web development trends, the discovery of data is in charge of automated agents, sophisticated search engines, and interoperability services. The era of Web Semantic allows machines the sharing and exploitation of knowledge, in a scalable and extensible way, where information is given a well-defined meaning. The structuring of information and the abstraction of intrinsic concepts constitute the formal knowledge, commonly known as ontology. Ontologies, or specifically, Web ontologies contribute to provide an adequate solution in knowledge representation. They enable the sharing of uniform structures for classifying knowledge regardless of the implementation language or the syntax used to represent it. However, as Web rapidly expands in size, the exigency of semantic organization beyond a fair arrangement of documents and text reveals the difficulty to glean knowledge from the Web, even though the state of art of natural language processing techniques or indexing processing is relevant. The distributed nature and the unknown reliability of knowledge on the Web need a clear conceptualization as well as an objective structuring for discovering unsuspected relationships and to summarize the data in novel ways that are machine understandable and useful to the data owner. This special issue introduces some novel approaches aimed at ontology-based knowledge discovery. Two papers address recommendation systems, proposing different perspectives of application domain and use. The paper by Morales-del-Castillo, Peis, Ruiz, and Herrera-Viedma presents a multiagent filtering and recommender system, which combines Semantic Web technologies and fuzzy linguistic modeling techniques to provide users valuable information about resources that fit their interests. The approach is designed to be used in the biomedical environments and provides an integrated solution to minimize the problem of access relevant information in vast document repositories. Pudota, Dattolo, Baruzzo, Ferrara, and Tasso introduce a novel unsupervised approach for recommending content-based tags by applying a combined set of techniques and tools that exploit tags, domain ontologies, and key phrase extraction methods. The approach achieves an ontology navigation, which enables the identification of meaningful ancestors for relevant extracted key phrases, to recommend significant metadata as new possible tags.