FAQ

1. What is Neuroph?
Neuroph is Java framework that can be used for creating neural networks. You can use it to easily add neural networks to your programs, or just to experiment with neural networks.

2. What are the requirements for Neuroph?
To use Neuroph you just need to have Java 1.6

3. How do I create and train neural network with Neuroph?
You can create networks with GUI editor easyNeurons using network specific wizards (see Networks in the main menu) , or in Java code simple as NeuralNetwork nnet = new Perceptron(2, 1); After that you have to provide training data as instance of TrainingSet class, and then to start the learning procedure nnet.learn(trainingSet); In GUI you can enter training set using the training set editor, and then to use the 'train' button.

4. For what neural networks can be used for?
Neural networks can be used for various tasks like classification, recognition, prediction, function approximation, and modelling.
Typical tasks are image, fingerprint or face recognition, stock prediction, pattern classification etc.

5. What types of neural networks Neuroph supports?
At the moment Neuroph supports the following networks: Adaline, Perceptron, Multi Layer Perceptron with Backpropagation, Hopfield, Konhonen, Hebbian, Maxnet, Competitive, Instar, Outstar, BAM, RBF network and NFR

6. How can extend the Neuroph to create new neural network architectures?
Derive your specific arhitecture fro NeuralNetwork class and provide method for network creation. After that create learning rule by deriving from some LearningRule (SupervisedLearning or UnsuprvisedLearning ). Add that learning rule to network during creation.

7. How can I contribute?
Send us your code if you have developed something using Neuroph, or choose something to do from our TODO list.