What it does
BCDn is able to determine whether or not a breast cancer tumor is malignant without an invasive biopsy. The model requires 20 data points that describe the tumor and a diagnosis with a 99 percent accuracy is returned. Our user interface allows individuals to input the data and get a result along with links to help them understand the diagnosis.
How we built it
Our model was built in Jupyter notebooks and our web app was built with HTML/CSS in flask.
The model was trained with a breast cancer dataset from the University of Wisconsin and tested with 12 different ANN models with variations of the data and fitted models. Our best model used Logistic Regression trained with a standardized version of the data.
Challenges we ran into
Finding new ways to feature engineer the data to get the best outcome. Connecting our model to our user interface.
Log in or sign up for Devpost to join the conversation.