Semantically enhanced search system for historical medical archives

Large-scale efforts to digitise historical documents are making it increasingly easy for researchers of history to carry out searches over vast amounts of historical data from their computers. Although the constant growth in the volume of digitised historical text is enriching the body of knowledge that scholars of history have at their fingertips, it can often be difficult to explore such data collections efficiently without becoming overwhelmed. Standard keyword-based search systems treat documents as collections of unrelated words, and do not take into account their structure and meaning. Accordingly, keyword searches will often return many irrelevant documents. Equally, shifts in terminology usage over time can make it difficult to formulate queries that will retrieve all relevant documents from long-spanning historical archives. In this paper, we describe a new semantically oriented system for searching archives of historical medical documents covering wide time spans. By applying text mining techniques to the archives, the system allows for efficient searching, firstly by automatically suggesting ways to expand queries with (possibly time-sensitive) related terms, and secondly by allowing search results to be refined/explored using medically and historically relevant semantic information.