Ontology-based Health Information search: Application to the neurological disease domain

The amount of information on health subjects on the Web and users´ interest in having efficient access to it continues to increase. This information can also be very useful for professionals, researchers and students in this field. But the usual word-based search does not allow full exploration of the information, pointing to the need for more semantic approaches. Aiming to contribute to providing a semantic search for information on Health, available using written Portuguese, the article presents enhanced search platform architecture based on the enterprise search platform (FAST), providing a processing pipeline capable of exploring annotation of entities and relations defined in a domain. A modular architecture was defined in which an ontology-based knowledge base has an essential role. As it is essential to make the user task as simple as possible, the platform also includes ontology navigation in the classes and relations of the entity to guide queries; use of the most frequent domain classes in the documents relating to a search to provide additional ways of navigation; and inclusion in the score of the ontology-guided annotations. As a first proof of concept, the architecture was instantiated in a specific domain, neurological diseases, for which an ontology was created.