NLP techniques fall into two broad classes: statistical and symbolic. Statistical techniques use information from the frequency distribution of words within a text to classify or extract information. Symbolic techniques use information from the structure of the language (syntax) and the domain of interest (semantics) to interpret the text to the extent necessary for encoding the text into targeted categories. Although some NLP applications exclusively use one or the other technique, many applications use both statistical and symbolic techniques. In this section, we give a brief background of NLP research in the medical domain and describe some statistical and symbolic NLP techniques used for classifying, extracting, and encoding information from biomedical texts, focusing on techniques useful for addressing the linguistic characteristics of patient medical reports described in the previous section.
Was this article helpful?