Since there are many coding patterns that occur very often when you use artificial neural networks, it is sometimes helpful to have shortcuts.

Here is an overview over these shortcuts. They are available only if either OpenANN/OpenANN or OpenANN/Convenience.h are included. In addition, there are some other helper functions in other headers that are not documented here. These headers are OpenANN/Preprocessing.h and OpenANN/Evaluation.h.

Function	Usage	Description
OpenANN::makeMLNN	OpenANN::Net net; OpenANN::makeMLNN(net, OpenANN::LOGISTIC, OpenANN::LOGISTIC, D, F, 2, 3, 3);	Create a multilayer neural network with D inputs, F outputs and two hidden layers with 3 nodes each. Note that you cannot set the standard deviation of the initial random weights or the bias of the MLNN.
OpenANN::train	OpenANN::Net net; // add layers OpenANN::StoppingCriteria stop; stop.maximalIterations = 100; OpenANN::train(net, "LMA", OpenANN::SSE, stop);	Trains a multilayer feedforward neural network. Note that you cannot pass any arguments to the optimization algorithm, that is you have to rely on the default settings for e.g. MBSGD which is usually not a good idea.