We started off wanting to tackle world issues, preferably with tech we're proficient with (machine learning). With an overarching interest in healthcare, we began looking at the process of diagnosing diseases and prescribing medicine. After abandoning the idea of precision medicine, we found our idea after reading research papers of medical inefficiencies.
What it does
Our machine learning model takes in biomarker data for cancer antigens as inputs, and predicts your chances of having pancreatic cancer. This can be used to either validate a doctor's opinion quickly and easily.
How I built it
We built our model in the Python language, with the Keras library with TensorFlow as a backend. This was done in Google Colab and Sublime Text for building and iteratively optimizing our model.
Challenges I ran into
It took us a while to overcome some of the errors we kept getting, but after persistence and talking to others we got our network up and running.
Accomplishments that I'm proud of
We're proud of how successful our model is at diagnosis compared to existing frameworks, and its potential benefit to society if deployed.
What I learned
I learned a ton about using pandas to read tables and other forms ofnumeric data for artificial intelligence.
What's next for Machine Learning to Detect Pancreatic Cancer with Biomarkers
We'd like to talk to experts in the industry and get our idea professionally validated!