$7\mathrm{.}4($SVMs without bias$)$ In the linear SVM formulation, we assume a classification boundary of the form

w

T

x$+$b

$,$ which includes a bias term

b in R

$.$ However, it is also possible to learn a linear SVM classifier without the bias term$-$i$.$e$.,$ using

w

T

x

as the classification boundary.