Activation function: Difference between revisions
From Rice Wiki
(Created page with "The '''activation function''' determines the output of a neuron in a neural network. It generates the output of the neuron from the linear combination of inputs calculated by the neuron. The choice of activation function varies depending on the machine learning task. A simple linear activation function or no activation function is just linear regression. A sigmoid function can be used for a classification task.") |
No edit summary |
||
| Line 2: | Line 2: | ||
The choice of activation function varies depending on the machine learning task. A simple linear activation function or no activation function is just [[linear regression]]. A sigmoid function can be used for a classification task. | The choice of activation function varies depending on the machine learning task. A simple linear activation function or no activation function is just [[linear regression]]. A sigmoid function can be used for a classification task. | ||
[[Category:Machine Learning]] | |||
Latest revision as of 00:15, 1 May 2024
The activation function determines the output of a neuron in a neural network. It generates the output of the neuron from the linear combination of inputs calculated by the neuron.
The choice of activation function varies depending on the machine learning task. A simple linear activation function or no activation function is just linear regression. A sigmoid function can be used for a classification task.
