The simplest case of an ANN is the perceptron model, illustrated in figure 2.5. If
we particularize the activation function ϕ to be the simple Threshold Function, the
output is obtained by summing up each of its input value according to the weights
of its links and comparing its output against some threshold θk. The output function
can be expressed using Eq. 2.7. The perceptron model is a linear classifier that has
a simple and efficient learning algorithm. But, besides the simple Threshold Function used in the Perceptron model, there are several other common choices for the
activation function such as sigmoid, tanh, or step functions.
The simplest case of an ANN is the perceptron model, illustrated in figure 2.5. If
we particularize the activation function ϕ to be the simple Threshold Function, the
output is obtained by summing up each of its input value according to the weights
of its links and comparing its output against some threshold θk. The output function
can be expressed using Eq. 2.7. The perceptron model is a linear classifier that has
a simple and efficient learning algorithm. But, besides the simple Threshold Function used in the Perceptron model, there are several other common choices for the
activation function such as sigmoid, tanh, or step functions.