Example-Neural-Network-Perceptron

Example-Neural-Network-Perceptron is an implementation of perceptron algorithm for supervised learning of binary classifiers, written in Java.

Junior project in class "Neural Network". Last updated 10/15/2014.

How to use

Select input file from dataSet directory.
Field 學習率 (learning rate) and 學習次數 (count of learning loop) columns or Field blank to use the default value.
Click 執行 (execute) button to run.

Feature

Graphics

Blue dots for testing set.

Red dots for training set.

Yellow dots for weight vectors.

DataSet (Trainging Set, Testing Set)

// src/neural_network_perceptron/Perceptron.java
...
int training_set_length = (int)Math.ceil(dataSet.size() * 2 / 3);
int testing_set_length = dataSet.size() - training_set_length;

Rate

// src/neural_network_perceptron/Perceptron.java
...
correctRateForTraining = (double)num_of_correct / (double)num_of_execute;
...
correctRateForTesting = (double)num_of_correct / (double)num_of_execute;

Default Setting

// src/neural_network_perceptron/Perceptron.java
...
learningRate = 0.8;
learningCount = 10000;

Dependency

JDK

$ java -version

java version "1.8.0_66"
Java(TM) SE Runtime Environment (build 1.8.0_66-b17)
Java HotSpot(TM) 64-Bit Server VM (build 25.66-b17, mixed mode)

Build and run

Compile

$ javac -d bin -sourcepath src src/neural_network_perceptron/*.java

Run

$ java -cp bin neural_network_perceptron.Main

Create jar file

$ jar cfe neural_network_perceptron.jar neural_network_perceptron.Main -C bin/ .

$ jar tf neural_network_perceptron.jar # list table of contents for archive

META-INF/
META-INF/MANIFEST.MF
.gitkeep
neural_network_perceptron/
neural_network_perceptron/Coordinate.class
neural_network_perceptron/FileData.class
neural_network_perceptron/Framework$CustomActionListener.class
neural_network_perceptron/Framework.class
neural_network_perceptron/Main.class
neural_network_perceptron/Perceptron.class