Automatic Translation to Controlled Medical Vocabularies

In the medical domain, over the centuries several controlled vocabularies have emerged with the goal of mapping semantically equivalent terms such as fever, pyrexia, hyperthermia, and febrile on the same (numerical) value. Translating unstructured natural language texts or verbatims produced by healthcare professionals to categories defined by a controlled vocabulary is a hard problem, mostly solved by employing human coders trained both in medicine and in the details of the classification system. In this chapter we survey the automatic translation or autocoding systems currently in use.