Summary of Linguistic Issues

We have described some of the linguistic characteristics of the sublanguage of patient medical records, including linguistic variation, polysemy, negation, contextual information, finding validation, implication, and coreference. If we want to automatically determine an individual patient's values for the variables used in our expert system, we must address these linguistic characteristics, using the types of information a physician uses to understand the meaning of the words and sentences in the reports. Below we describe some of the techniques current natural language processing research employs for extracting information from clinical texts.

0 0

Post a comment