Files

learningWithBackpropagation

Quinn Liu

added instructions for how to use code

Jan 11, 2014

35daa9e · Jan 11, 2014

This branch is 1 commit ahead of joemehr/MachineLearning:master.

Name		Name	Last commit message	Last commit date
parent directory ..
README.md		README.md	added instructions for how to use code	Jan 11, 2014
checkNNGradients.m		checkNNGradients.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
computeNumericalGradient.m		computeNumericalGradient.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
debugInitializeWeights.m		debugInitializeWeights.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
displayData.m		displayData.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
ex4.m		ex4.m	implemented feedforward propagation and cost function for neural network	Jan 11, 2014
ex4data1.mat		ex4data1.mat	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
ex4weights.mat		ex4weights.mat	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
fmincg.m		fmincg.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
nnCostFunction.m		nnCostFunction.m	implemented neural network backpropagation and regularized gradient	Jan 11, 2014
predict.m		predict.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
randInitializeWeights.m		randInitializeWeights.m	implemented neural network backpropagation and regularized gradient	Jan 11, 2014
sigmoid.m		sigmoid.m	added new folder for showing the backpropagation neural network learn…	Jan 10, 2014
sigmoidGradient.m		sigmoidGradient.m	implemented derivative of sigmoid function also known as sigmoid grad…	Jan 11, 2014

README.md

Neural Network Learning with Backpropagation

How to use this code:

1. install Octave or Matlab

fork this repo and clone it locally!
navigate into the folder with the above files
type ex4 to the command line to view a digit recogition neural network with about 95% success rate trained using backpropagation.