-
Notifications
You must be signed in to change notification settings - Fork 1
Neural Network
Jelle Spijker edited this page Jun 19, 2015
·
2 revisions
The used Neural Network is a a network of interconnected neurons, called a multi-layer perceptor MLP. This type of neural network has one input layer, one or more hidden layers and an output layer. Each neuron consist of a summation function that is fed into a sigmoid threshold function. If the threshold is reached the output together with a weight acts as a input for the following layer.
The current class defined in the SoilMath namespace, has to be created as an object. This object consist of a number of number of input layers with a bias neuron, one hidden layer with a bias neuron, and a output layer. For each layer an equal amount of weights are defined.
Copyright Jelle Spijker 2015 All Rights Reserved