GitHub - basicmachines/mlp: Python implementation of feed-forward multi-layer perceptron neural network using only numpy and scipy

mlpnet.py

Python implementation of feed-forward multi-layer perceptron (MLP) neural networks using numpy and scipy based on theory taught by Andrew Ng on coursera.org and adapted from the Octave code examples from this course as well as updates from the 2017 deeplearning.ai Deep Learning Specialization course..

This is still a work-in-progress. Still some bugs and fixes to do so don't recommend using this yet...

Applications

Classification
Prediction
Model Predictive Control (MPC)
Approximation of non-linear functions

Main Classes

MLPLayer - Multi-layer perceptron neural network layer class (used by MLPNetwork)
MLPNetwork - Multi-layer perceptron neural network class
MLPTrainingData - Object to store training data (inputs and outputs).

Other Functions

train - Used to handle training of a network.