DuckHacks Theme
Inspiration
We wanted to make a unique identifier for people based on their habits.
What it does
Uses soft margin support vector machines to cluster and classify people based on their habitual data
How I built it
Python Sklearn, youtube, trial and error with other neural networks
Challenges I ran into
Started with recurrent neural networks. Couldn't properly reshape time series into a needed matrix shape. Then realized SVM was actually better
Accomplishments that I'm proud of
Made this
What I learned
A lot about python libraries. Not a lot about math
What's next for noU
Write in lower level language, C++, for architecture optimization Include other neural networks for pattern recognition, that contribute to overarching architecture. This allows for better analysis on datatypes we haven't used yet (eg, image data, natural language processing, or denoising audio).
Log in or sign up for Devpost to join the conversation.