$2-$ Perceptron Algorithm $(40$P$)$

Consider the following dataset for a binary classification task:

Assume the perceptron starts with weights $w1=0\mathrm{.}5,w2=0\mathrm{.}8$ and bias $b=-0\mathrm{.}3.$ The learning rate is $=0\mathrm{.}4$

Perform three full passes $($epochs$)$ of the perceptron algorithm and update the weights and bias for each misclassified point. Choose one point from the red class, one point from the black class at each step in training.

Show the updated weights and bias after processing all points.

Draw the decision boundary.