Non supervised classification tools adapted to supervised classification

Let X be some set of individuals. We consider the following two mappings: $$\begin{array}{*{20}{c}} {R:X \to {R^p}} \\ {\Omega :X \to \left\{ {{\omega _1}, \ldots ,{\omega _n}} \right\}} \end{array}$$ For an individual x, x ∈ X, R(x) is its representation (R p being the feature-space) and Ω(x) is its class. A training set in a supervised classification problem consists in the data (R(x), Ω(x)) for any x belonging to a subset T of X.