i am looking for help in part A.

Thanks

selected by cross-validation reasonable? (b) By comparing the final test results, what type of kernel do you think is more appro- priate for the tremor dataset? 4. ,n) with z, e Rd and (10 pts) Let the training data be given by {(x,,w) : i = 1,2, Vi E {1). The standard SVM formulation (1) mentioned in Questions 2 and 3 is usually referred to as 1-norm soft margin SVM. A well-known variant modification of the standard SVM is so-called SVM with 2-norm soft margin which is given by mnw.he 11 wif + ,4 0, i=1, ,n. where C 0 is a trade-off parameter. Pick up one from the following questions and complete it. (a) (10 pts) Show that 2-norm soft margin SVM 3) is equivalent to the following ra min F(w, b) :=- max(0, 1 - (w'xi + b) Furthermore, prove that F is a convex function of (w,b). Now using the Lagrangian multiplier theorem, show that the dual formulation of the SVM with 2 norm soft margin is given by Then, using this dual formulation, describe how the above lincar 2-norm SVM can be extended it to the nonlinear case. selected by cross-validation reasonable? (b) By comparing the final test results, what type of kernel do you think is more appro- priate for the tremor dataset? 4. ,n) with z, e Rd and (10 pts) Let the training data be given by {(x,,w) : i = 1,2, Vi E {1). The standard SVM formulation (1) mentioned in Questions 2 and 3 is usually referred to as 1-norm soft margin SVM. A well-known variant modification of the standard SVM is so-called SVM with 2-norm soft margin which is given by mnw.he 11 wif + ,4 0, i=1, ,n. where C 0 is a trade-off parameter. Pick up one from the following questions and complete it. (a) (10 pts) Show that 2-norm soft margin SVM 3) is equivalent to the following ra min F(w, b) :=- max(0, 1 - (w'xi + b) Furthermore, prove that F is a convex function of (w,b). Now using the Lagrangian multiplier theorem, show that the dual formulation of the SVM with 2 norm soft margin is given by Then, using this dual formulation, describe how the above lincar 2-norm SVM can be extended it to the nonlinear case