Skip to content

TheHiddenLayer/perceptron

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Perceptron

In progress if I'm not busy with other things

A perceptron is just a recurrently-activated neural network that finds useful features or business intelligence in input data. Here's an insanely simple model:

You should ideally santize your data before using this API. You could use pandas or something similar to clean your data, not even breaking a sweat.

Ended up choosing to write the source code in Rust. fast and fun. I'm pretty busy with school and other things right now so you're heavily encouraged to contribute.

Keep in mind that this is a binary classifier so craft each line in your .csv in the format of <binary class [0 or 1]>, cartesian x, cartesian y.

After you've done all that (and once this project is complete) it gets damn simple: cargo run <path_to_.csv>

If you're interested in seeing it work on a small dataset

cargo run < examples/small.csv

With the addition of a visualizer, the goal is to be able to see the perceptron updates in real-time.

todo

  1. Finish training

  2. Create unit tests for the more nuanced datasets which may or may not be linearly separable.

Releases

No releases published

Packages

No packages published

Languages