The perceptron algorithm is a basic artificial neuron and is considered to be a building block of neural networks. It takes inputs, applies weights and produces output using an activation function. A supervised learning algorithm, it is a binary classifier, separating data points using a linear combination of the input variables.

The algorithm consists of

• an input vector of the form

• a weight vector

• a bias denoted as b
• the summation function z

• activation function (z)

The mathematical concepts needed in order to understand the perceptron algorithm are:

• definition of independent variables
• what a linear equation is
• vectors
• dot product
• heaviside step function

NOTE: The perceptron algorithm can classify OR and AND functions but not XOR as it is not linearly separable.