Q$.3(25$ marks$)$

The Perceptron is an artificial neural network $($ANN$)$ developed by Frank Rosenblatt in $1958.$

It is an algorithm for supervised learning of binary classifiers. The perceptron is a type of linear

classifier, i$.$e$.$ a classification algorithm that makes its prediction based on a linear predictor

function combining a set of weights with the input vector. Answer the following questions.

$($a$)$ Draw a Perceptron with $4$ inputs and $1$ output. Your answer should include labels of inputs

$x_i,$ output $y,$ weights $w_i$ and bias $b.$

$(2$ marks$)$

$($b$)$ Provide the learning rule that is used to adjust the weights and bias of the Perceptron.

$(3$ marks$)$

$($c$)$ You are required to train a Perceptron to classify the type of Iris plant by using the dataset

given in Table Q$\mathrm{.}3$a$.$ Choose initial weights $w_1=w_2=w_3=w_4=-0\mathrm{.}5,$ initial bias $b=$

$1,$ learning rate $=0\mathrm{.}25$ and threshold $=0\mathrm{.}1.$ Redraw and fill in Table $Q\mathrm{.}3b$ in your

answer booklet for the first epoch.

Table Q$\mathrm{.}3$a: Iris Dataset

Table Q$\mathrm{.}3$b: First Epoch

$($d$)$ Continuing from $Q\mathrm{.}3c,$ adjust the weights and bias for the second epoch. Redraw and fill in

Table Q$\mathrm{.}3$c in your answer booklet for the second epoch.

$(9$ marks$)$

Table Q$\mathrm{.}3$b: First Epoch

$($d$)$ Continuing from Q$\mathrm{.}3$c$,$ adjust the weights and bias for the second epoch. Redraw and fill in

Table Q$\mathrm{.}3$c in your answer booklet for the second epoch.