The main advantages of ANN are that – depending on the activation function
– they can perform non-linear classification tasks, and that, due to their parallel
nature, they can be efficient and even operate if part of the network fails. The main
disadvantage is that it is hard to come up with the ideal network topology for a
given problem and once the topology is decided this will act as a lower bound for
the classification error. ANN’s belong to the class of sub-symbolic classifiers, which
means that they provide no semantics for inferring knowledge – i.e. they promote a
kind of black-box approach.