Dimension-reduction methods aim at defining a score that maximizes signal diversity. Two approaches, namely maximum entropy weights and maximum variance weights, are provided.
