# Is the percepetron algorithm's convergence dependent on the linearity of the data?

Does the fact that I have linearly separable data or not impact the convergence of the perceptron algorithm? Is it always gonna converge if the data is linearly separable and not if it is not ? Is there a general rule ?