Difference between revisions of "Activation Function"
From Open Risk Manual
Wiki admin (talk | contribs) |
(No difference)
|
Latest revision as of 21:02, 11 September 2020
Definition
In the context of a neural network based Risk Model (such as the Perceptron), the activation function of a network node defines the output of that node given an input or set of inputs.
It is essentially a choice of functional form, which may have implications for the ease of implementation or performance of the model. It is a generalization of the concept of the distribution of the response variable in e.g. logistic regression versus probit regression models.