Neural networks
A neural network consists of the following information:
- A directed acyclic graph
where the source vertices are called the input layer and the sink vertices are called the output layer , and - For each
we have a perceptron . In the Modelling framework this would simulate something with
and . To run the model you compute a value for each vertex for . For input we associate to vertex . Then for all once all of have values we set To generate output vector . Normally the topology of
is chosen so that the vertices can be partitioned into for some . Where for all and . Where is the input layer and is the output layer all other layers are called hidden layers.