How about more concrete? It could be applied to a pattern recognition problem. For example, classifying email messages as spam and not-spam. The dots in 2-D

How about more concrete? It could be applied to a pattern recognition problem. For example, classifying email messages as spam and not-spam. The dots in 2-D space represent some characteristics of the message (a poor example would be number of words, average number of letters in each word).