What you're observing
A perceptron adjusts weights in response to misclassified samples. Learning rate and cluster noise influence convergence and mistakes.
w ← w + η(y−ŷ)x
Technology / Interactive Lab 04
Train a tiny neural classifier and watch its decision line gradually separate two noisy clusters.
A perceptron adjusts weights in response to misclassified samples. Learning rate and cluster noise influence convergence and mistakes.