Abstract: The availability of comprehensive electronic medical records that
include narrative reports provides an opportunity for natural language
processing (NLP) technologies to play a major role in clinical research.
One of the main advantages of employing these technologies is the
automatic extraction of relevant clinical information to identify
critical illness phenotypes and to facilitate clinical and translational
studies of large cohorts of critically ill patients. In this talk, I
will present an NLP system for the task of pneumonia identification.
Based on the information extracted from the narrative reports associated
with a patient admitted in the intensive care unit, the task is to
identify whether or not the patient is positive for pneumonia. I will
show that, an approach using statistical feature selection, in which
only a small subset of informative features from the initial feature
space is considered, achieves significantly better results than a
baseline, which uses all the features from the same feature space. The
addition of a feature that extracts the assertion value of all pneumonia
expressions from the clinical dataset considered further improves the
performance of the NLP system for this task.

Speaker Bio:
Cosmin Adrian Bejan is a senior fellow in the Division of Biomedical and
Health Informatics at the University of Washington. Prior to his current
Technologies at the University of Southern California. Cosmin received
his M.S. and B.S. degrees in computer science from the University of
Iasi, Romania. He holds a Ph.D. degree in computer science from the
University of Texas at Dallas, where he investigated natural language
processing and machine learning methodologies in order to capture the
semantics of the event structures that are encoded in text. His research
interests are in the areas of natural language processing, biomedical
informatics, and machine learning with a focus on event semantics,
open-domain and clinical information extraction, and commonsense causal
reasoning.