Potentially Explaining Literature Based Discoveries in Non-Medical Domains

Literature base discovery has been used for many years as a way to identify possible discoveries of related concepts where the concepts have never been mentioned previously in the same document. Linking concepts and the nature of the relationship between the hidden discovery and the linking concepts may be gleaned from the documents using existing relationship extraction techniques. Little has been done to automatically explain the relationship between the hidden knowledge the literature based discovery candidate pair. This paper presents our work in automatically identifying the nature of the relationship of the hidden discovery using supervised machine learning techniques. It starts with the research that has been completed on the medical domain and proposes work that may be done in the intelligence domain.