https://github.com/johnberroa/perceptrons

My implementations of perceptrons and MLPs
https://github.com/johnberroa/perceptrons

neural-network perceptrons

Last synced: 7 months ago
JSON representation

My implementations of perceptrons and MLPs

• Host: GitHub
• URL: https://github.com/johnberroa/perceptrons
• Owner: johnberroa
• Created: 2017-11-18T22:58:29.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2018-03-11T13:21:06.000Z (over 8 years ago)
• Last Synced: 2024-12-28T05:43:02.173Z (over 1 year ago)
• Topics: neural-network, perceptrons
• Language: Python
• Size: 152 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Overview

#### Note: MLP still W.I.P.

Three implementations of perceptrons:

* The *Logical Perceptron* creates a logical operator learning perceptron, i.e. 

it learns how to compute "and", "or", "nand", and "nor".  It returns its weights and test sample performance so that training

can be seen quantitatively.  For more understanding, it also plots the movement of the decision boundary that linearly separates

the data.  

* The *Cluster Perceptron* seperates two clusters of data.  It also prints out the weights and the results of the training on

both the training and test sets.  In theory, one can input their own cluster data and it will work out of the box.  

* The *Multilayer Perceptron*, better known as a Neural Network, can be built to any desired network size.  The activation functions, weight initializations, and learning rates are customizable at the layer level.  As always, graphs will be included.  

Below are pictures showing prototypical output from the algorithms:

![Logic](/gfx/logicperceptron.png?raw=true "Logic Perceptron")

![Cluster](/gfx/clusterperceptron.png?raw=true "Cluster Perceptron")