I don't know how to solve this

(2) Suppose you are given a dataset of as, y pairs {(331, yi)};N=1, with 56.; E X and yi 6 {i1}. Assume that 71+, n_ of the m have label +1, 1 respectively (so n+ + m = N) and also assume you are given a kernel K and an associated feature map t1> : X > .7: to some Hilbert space .7: so KW: 93') = (M03), (1)0937);- Derive a classication rule, using kernel products and the sign function, that assigns to a new test point the label of the class whose mean is closest in feature space. Note: pay careful attention to the requirements in boldface; in particular, nothing tells you you can take a mean in the space X. Also, a hint: you may nd squared distances convenient to work with