Skip to main content

Castine Melissa

Description

Natural language processing algorithms that accurately screen clinical documents for suspected pneumonia must extract and reason about whether these mentions provide evidence that supports, refutes, or represents uncertainty. Our efforts extend existing algorithms [1] and taxonomies [2] that can be leveraged by NLP tools for more accurate handling of uncertainty for suspected pneumonia case review.

Objective

We sought to classify evidence that supports, refutes, or contributes uncertainty when reviewing cases of suspected pneumonia. We extend an existing taxonomy of uncertainty to classify these phenomena with the goal of improving existing Natural Language Processing (NLP) algorithms.

Submitted by Magou on