Scott Leroy Duvall PhD, Presenter: Research Grant, Amgen Inc Research Grant, Anolinx LLC Research Grant, AstraZeneca PLC Research Grant, F. Hoffmann-La Roche Ltd Research Grant, Merck & Co, Inc Research Grant, Mylan Inc Research Grant, PAREXEL International Corporation Research Grant, Shire plc

1) Review information extraction methods for building rule-based, grammar-based, and machine-learning NLP systems with examples of when to use each. 2) Demonstrate the creation of manually created reference standards against which to measure NLP systems. 3) Present a survey of open-source tools for NLP and manual chart review and how these can be built upon.

Natural language processing (NLP) is a term that describes a range of techniques for identifying, understanding, and analyzing information from text. Some of the earliest applications of NLP in medicine were on imaging reports. Attendees will be walked through both simple and complex NLP methods with examples of how and when they are best used in imaging. Several open-source tools will be demonstrated with information provided on how these tools can easily be built upon for customized needs.

Duvall, S, So Many to Choose: An Overview of Natural Language Processing Methodologies and When to Use Each  .  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14002371.html