A classifier uses four linear discriminant functions in the plane: g1 = 1, 92 = 2, 93 = 22 +1-3, and 94 = X2 - X1 +1. i) Sketch the line boundaries in pattern space and label each region with a 4-bit binary codeword. ii) Make a logical table, identifying for each codeword whether or not there is an associated region. iii) Sketch the four-dimensional hypercube in state space. (Don't panic ! Use two 3- dimensional projections, one yielding a 3D cube for the case 71 = 0 and the other giving a 3D cube for 71 = 1.). Label each vertex with its corresponding binary codeword. Draw a curved line joining each vertex in the first cube with the corresponding vertex in the second cube. ii) Consider the discriminant function y = sgn( _ T; - 2.5). Sketch the decision region induced i=1 by this rule in the (2D) feature space