Exploring and exploiting knowledge in multiple resources

With the huge growth of the Internet and the WWW, several scientific fields have witnessed a proliferation of information available online. Techniques that are able to extract relations and associations that may exist in various resources including structured (ontologies and taxonomies) and unstructured (corpora) to inform search and can offer enhanced recommendations are highly desirable. The aim of this work is to pull together multiple bioinformatics resources with different structures such as ontologies and corpora and develop reasoning methods to extract semantic relations and hidden associations, which do not exist in the original resources. We have designed methods which reason through these resources and contain semantic rules (for instance, identifying sibling relations) that allow us to infer more hidden information between resources and aim to enhance recommendations in situations as described above. Moreover, we have designed a recommender approach for the bioinformatics field that provides recommendations on content (i.e. articles) as a personalised service for users based on information extracted from multiple resources as well as their individual profiles.