Final answer:
The true statement for k-NN classifiers is that k-NN does not require an explicit training step. Larger or smaller values of k affect the smoothness and distinctness of class boundaries differently, and the decision boundary is not necessarily linear but can be complex.
Step-by-step explanation:
The k-NN classifiers (k-Nearest Neighbors) is a type of instance-based learning, or lazy learning, where the function is only approximated locally and all computation is deferred until function evaluation. The correct statement regarding k-NN classifiers is b) k-NN does not require an explicit training step.
For k-NN classifiers, the classification accuracy is not necessarily better with larger values of k. Instead, the best choice of k depends on the data; larger values of k can reduce the effect of noise on the classification, but can also make the boundaries between classes less distinct. Concerning the decision boundary, it is actually smoother with larger values of k, contrary to what's suggested in the c) option. Lastly, the decision boundary of k-NN is not linear (option d), it can be very complex and is non-parametric.