CADET vs. k-Nearest

CADET represents instances with rich symbolic descriptions so it needs a new similarity metric (e.g., size of largest shared subgraph).
Both can use multiple examples in classification, but they combine them in different ways. CADET must have special rules for combining.
CBR, in general, shows a tight coupling between case retrieval, knowledge-based reasoning, and problem solving.
CBR is useful for simple matching of queries such as help-desk questions.