Which of the following is not a Clustering method?

Explanation

Clustering is unsupervised: it groups unlabelled data by similarity. K-means partitions data into k clusters around centroids, Kohonen's self-organising feature map produces a topology-preserving unsupervised grouping, and the agglomerative method is the bottom-up form of hierarchical clustering. K-nearest neighbour is different in kind — it is a supervised lazy classification algorithm that assigns a label to a new point from the labels of its k closest training examples, so it needs labelled data and produces a classification, not clusters. Option 3 is therefore the answer.