Making tasks easier for patients and saving time for doctors and nurses
What it does
The chatbot understands user's words and queries to handle tasks such as retrieving lab reports, schedule appointments, refill prescriptions, send messages to doctors, and retrieve billing information.
How we built it
We used Chatito to generate the data and used the Rasa framework to handle the chatbot functionality and the NLU component.
Challenges we ran into
-Generating good data is difficult -New to natural language processing -Trail and error trying to increase accuracy -Getting the chatbot to run on everyone's computer took a long time. There was some issues with versions and weird configuration settings we had to do that required a lot of research.
What we learned
How to make a basic chatbot Precision, recall, and f1-score We learned the importance of having good data. Data that doesn't have enough variety could lead to over-fitting.