In the previous article, Machine Learning Basics and Perceptron Learning Algorithm, the assumption was that the Iris Data Set trained by Perceptron Learning Algorithm is linear separable, so the number of misclassification on each training iteration eventually converge to 0.
Submit an article or tip