Let (x1,y1), (2, y2), (N, YN) be N points on which we find a SVM classifier...

  

Transcribed Image Text:

Let (x1,y1), (2, y2), (N, YN) be N points on which we find a SVM classifier using =N 02-2022da04 dual SVM formulation which is g maximize a aα YiYjx.xven below: i=1 - i=N j=N i=1 subject to 1-2003/0410 1002200 i=1 2023/04/02-26 Let Oд be the value of the objective function at the optimal solution returned by the dual SVM formulation for this problem. Now we add (N+1, a new pointhe dua) and find a SVM classifier by solving the dual formulation again. Let OB be the value of the objective function at the optimal solution for this problem. Considering the following three relationships (a) OA > OB, (b) OA=0B, (c) OA < Og determine which of these relationships are possible and give a mathematical argument for your answer in each case. 82 give 002 Let (x1,y1), (2, y2), (N, YN) be N points on which we find a SVM classifier using =N 02-2022da04 dual SVM formulation which is g maximize a aα YiYjx.xven below: i=1 - i=N j=N i=1 subject to 1-2003/0410 1002200 i=1 2023/04/02-26 Let Oд be the value of the objective function at the optimal solution returned by the dual SVM formulation for this problem. Now we add (N+1, a new pointhe dua) and find a SVM classifier by solving the dual formulation again. Let OB be the value of the objective function at the optimal solution for this problem. Considering the following three relationships (a) OA > OB, (b) OA=0B, (c) OA < Og determine which of these relationships are possible and give a mathematical argument for your answer in each case. 82 give 002