What's this about?

Classify by finding the most similar examples in your dataset, then voting based on their labels.