Add to Calendar
When:
June 8, 2017 @ 12:00 pm – 1:00 pm
2017-06-08T12:00:00-07:00
2017-06-08T13:00:00-07:00
Where:
MSOB, Conference Room X-275
1265 Welch Rd
Stanford, CA 94305
USA
1265 Welch Rd
Stanford, CA 94305
USA
Cost:
Free
Contact:
Marta Vitale-Soto
Ken Jung,
Research Scientist,
BMIR, Stanford University
ABSTRACT:
Longitudinal clinical data captured in Electronic Health Records are complex and high dimensional. Researchers must make challenging decisions that can have a significant impact on predictive performance and clinical findings when preparing the data for analysis. Finding a good representation for clinical data is a critical step in designing studies and constitutes a grand challenge for machine learning researchers interested in health care. We present work that focuses on how to learn good representations of both structured data in the form of diagnosis, medication and procedure codes along with unstructured data in the form of free text clinical notes.