GitHub - druuu/letterRecognitionAnn: Artificial Neural Networks for letter recognition using python and octave.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
.letterRecog.py.swp		.letterRecog.py.swp
README.txt		README.txt
ex4.m		ex4.m
ex4data1.mat		ex4data1.mat
ex4weights.mat		ex4weights.mat
fmincg.m		fmincg.m
grads.mat		grads.mat
img		img
letRec.sh		letRec.sh
letRecGui.py		letRecGui.py
letterRecog.py		letterRecog.py
my_drawing.jpg		my_drawing.jpg
nnCostFunction.m		nnCostFunction.m
predict.m		predict.m
randInitializeWeights.m		randInitializeWeights.m
resize.py		resize.py
result.txt		result.txt
sigmoid.m		sigmoid.m
sigmoidGradient.m		sigmoidGradient.m

Repository files navigation

Artificial Neural Networks being used. Core code is in octave which has been
taken from coursera.org machine learning course and modified.
The training data is set of fonts with different distortions.

1.run letRecGui.py musing python2.7
2.ex4.m is main file of neural nets which is called by letRecGui.m
3.fmicg is cost minimization function file
4.to train neural nets change maxiter in ex4.m file and comment
load('grads.mat') line