Question 4 (10 points total) Select the most appropriate ans…
Question 4 (10 points total) Select the most appropriate answer: Part A (5 points) A K-nearest neighbor model with K = 5 (number of neighbors) i) classifies based on 5 training instances ii) requires the training data to be available for inference iii) is a lazy learner iv) all of the above Part B (5 points) A K-nearest neighbor model with Gaussian kernel g(xi , x0; λ) = exp(−0.5∥xi − x0∥2/λ) predicts which class as the λ parameter in the kernel is increased without bound. i) class of the instance nearest to the target instance x0 ii) cannot be determined iii) minority class (least frequent class) iv) majority class (most frequent class)
Read Details