Illness Diagnosis Assistant

Inspiration

Our team wanted to build something that everyone could use for their own benefit. We thought that building an illness diagnosis assistant was the perfect project because everyone gets sick here and there, but sometimes it is difficult to figure out what you have and how to treat your illness properly, so we thought we could use code to help individuals narrow their disease and treatment options down based off of their symptoms and age. Not only can it help people out, the app could also be used to track symptoms and narrow down possible illness for hospitals. This could then be used to organize patients by their urgency and make scheduling more efficient.

What it does

Our app takes symptoms and an age value from user input and then uses the OpenAI API to return a list of potential diseases the user may have along with simple treatment options.

How we built it

We build this program using the OpenAI API and training the AI to return a list of diseases along with treatment options in response to the user inputting a list of symptoms, their age, and asking the AI what their diagnosis is. We created a python program that prompted the user for age and symptom information and stored that data in string variables. We ensured that the question that OpenAI was asked always included that stored string data. Finally, we used TKinter to design the GUI.

Challenges we ran into

Halfway through the project, two of our teammates dropped, so that stopped us from our original idea of making a neural network to predict illness. The remaining members did not know about neural networks and we couldn't find a useful dataset, so we decided to focus on using OpenAI.

Accomplishments that we're proud of

Our OpenAI prompt does a pretty good job at predicting illness and suggesting relief for the symptoms.

What we learned

For the three of us, this was our first project relating to AI and Machine Learning, so it was a starting experience in the field. We also saw the power of OpenAI and were very surprised by it.