Prokinopub
Abstract
Searching through scientific articles for relevant information has proven to be a difficult and time-consuming task for researchers. Many improvements to the traditional, keyword-based search method have been proposed. While several of them have used Natural Language Processing techniques or other computationally expensive methods for retrieving useful information, such systems have been shown to not be very usable. Many modern systems have focused on creating better and more intuitive user interfaces and enhanced user experiences. In this thesis, we have developed a new ontology-driven scientific document mining system that aids the user in formulating a document-matching request. A suitable ontology browser helps the user by suggesting how terms and concepts are related to each other. Not only can the user browse the ontology, but (s)he can simultaneously formulate the search request. Our prototype implementation works in the Protein Kinase domain, but it can be applied to other domains in the future.