More often than not, the analyst wants to find the pair of observations that is closest in terms of certain metrics, such as Euclidean distance. For example, on the popular SAS-L archive group, Randall Powers posted such a question at here, but he …
Tag: proc modeclus
Using SAS to find the best k for k-Nearest Neighbor classification
Least-square (regression) and nearest-neighbor are the most fundamental methodologies for supervised classification [Ref. 1]. Even though they are pretty old, they are still popular and widely used in academia and industry. There is a trade-off in comp…