Simple 3 layer backpropagation neural network w/3 demo applications (point classifier, XOR, 0-9 recognition)
Java Groovy
Switch branches/tags
Nothing to show
Latest commit df5ee93 Apr 21, 2014 @guycole README cleanup
Permalink
Failed to load latest commit information.
.gradle/1.11/taskArtifacts
.idea
.settings
demo1
demo2
demo3 latest Apr 13, 2014
network
.classpath
.gitignore
.project first candidate Sep 13, 2012
BackProp1.iml
README.md README cleanup Apr 21, 2014
backProp3.png
build.gradle
settings.gradle

README.md

BackProp1

Simple 3 layer backpropagation neural network w/3 demo applications (point classifier, XOR, 0-9 recognition)

Also available at http://backprop1.sourceforge.net/

Update: 12 April 2014

BackProp1 now uses gradle.

Original BackProp1 on eclipse moved to "original" branch.

"gradle runDemo1" generates training set and trains to solve point classification for the equation "y = -5x+2" (you can view the training set w/gnuplot by using demo1/src/main/script/gnuplot1.dem)

There is no gradle support for demo2, an XOR demo.

demo3 screenshot

"gradle runDemo3" is an interactive digit recognizer. It trains for the digits 0 through 9, and then allows one to change the pattern values then submit for classification.