Q$2-$ We are given $10$ feature vectors that originate from two classes $W_1$ and $W_2$ as follows

$W_1$:$[0\mathrm{.}1,-0\mathrm{.}2{]}^T,[0\mathrm{.}2,0\mathrm{.}1{]}^T,[-0\mathrm{.}15,0\mathrm{.}2{]}^T,[1\mathrm{.}1,0\mathrm{.}8{]}^T,[1\mathrm{.}2,1\mathrm{.}1{]}^T$

$W_2$:$[1\mathrm{.}1,-0\mathrm{.}1{]}^T,[1\mathrm{.}25,0\mathrm{.}15{]}^T,[0\mathrm{.}9,0\mathrm{.}1{]}^T,[0\mathrm{.}1,1\mathrm{.}2{]}^T,[0\mathrm{.}2,0\mathrm{.}9{]}^T$

Check whether these are linearly separable and, if not, design an appropriate multilayer perceptron with nodes having step function activation to classify the vectors in the two classes. The multilayer perceptron must be done manually with mathematical expressions.