When updating the weight vector of the Perceptron, the algorithm given in class chooses one of the miss$-$

classified points randomly for the update. Another method to update the Perceptron weight vector is to

choose the miss$-$classified point that has the highest error. Apply this method to the dataset given

below until convergence, where the perceptron update rule takes the form w$^((\backslash$tau $+1))=$w$^((\backslash$tau $))+\backslash$eta x$\_($n$)$t$\_($n$).$ The

error function of the perceptron is $-$w$^($T$)$x$\_($n$)$t$\_($n$).$ Show the decision boundary obtained after each iteration

until all points are correctly classified. Assume that the weight vector is initialized as w$^((0))=[[1],[0]]($there is no

bias$).$ Use learning rate parameter $\backslash$eta $=0\mathrm{.}5.$

For class x$,$t$=1$

For class O$,$t$=-1$