A simple neural network written in Java.
This repository contains a neural network with backpropagation.
You train a neural network that has two input neurons, four hidden neurons and one output neuron.
The input is represented by two numbers between 0 and 1. The expected output should be the addition of the two numbers.
Example:
Data which the network was trained with:
input one | input two | expected output | calculated output |
---|---|---|---|
0.2 | 0.2 | 0.4 | 0.4002 |
0.5 | 0.3 | 0.8 | 0.7998 |
0.2 | 0.7 | 0.9 | 0.8999 |
Data calculated by the neural network:
input one | input two | expected output | calculated output |
---|---|---|---|
0.4 | 0.1 | 0.5 | 0.5030 |
0.6 | 0.2 | 0.8 | 0.7975 |
0.3 | 0.4 | 0.7 | 0.7067 |