Intuition for perceptron weight update rule

Question

I am having trouble understanding the weight update rule for perceptrons:

w(t + 1) = w(t) + y(t)x(t).

Assume we have a linearly separable data set.

Answer 1

A better derivation of the perceptron update rule is documented here and here. The derivation is using gradient descent.

• Basic premise of gradient descent algorithm is find the error of classification and make your parameters so that error is minimized.

PS: I was trying very hard to get the intuition on why would someone multiply x and y to derive the update for w. Because w is the slope for a single dimension (y = wx+c) and slope w = (y/x) and not y * x.