is a type of artificial neural network that consists of a single layer of output neurons, with each input connected to every neuron in the output layer. It is primarily used for binary classification problems.
- Input Layer: The perceptron takes a set of inputs, which are typically binary or real-valued.
- Weights: Each input is connected to the output neuron by a weight
- Bias: There is also a bias term b added to the weighted sum to allow for the shifting of the decision boundary.
- Activation Function: The perceptron uses a simple activation function to decide whether the neuron fires (outputs 1) or not (outputs 0). Usually, a step function (Heaviside step function) is used.