Too many impersonal doctor's office experiences, combined with the love of technology and a desire to aid the healthcare industry.
What it does
Takes a conversation between a patient and a doctor and analyzes all symptoms mentioned in the conversation to improve diagnosis. Ensures the doctor will not have to transcribe the interaction and can focus on the patient for more accurate, timely and personal care.
How we built it
Ruby on Rails for the structure with a little bit of React. Bayesian Classification procedures for the natural language processing.
Challenges we ran into
Working in a noisy environment was difficult considering the audio data that we needed to process repeatedly to test our project.
Accomplishments that we're proud of
Getting keywords, including negatives, to match up in our natural language processor.
What we learned
How difficult natural language processing is and all of the minute challenges with a machine understanding humans.
What's next for Pegasus
Turning it into a virtual doctor that can predict illnesses using machine learning and experience with human doctors.