If the Bayesian decision boundary is highly nonlinear we should have a low value of k since that increases flexibility and hence creates a more non-linear boundary, in the answer a high value of k is mentioned which will produce a highly smooth(linear) boundary
If the Bayesian decision boundary is highly nonlinear we should have a low value of k since that increases flexibility and hence creates a more non-linear boundary, in the answer a high value of k is mentioned which will produce a highly smooth(linear) boundary