A monotonic measure for optimal feature selection

Abstract

Feature selection is a problem of choosing a subset of relevant features. In general,only exhaustive search can bring about the optimal subset. With a monotonic measure, exhaustive search can be avoided without sacrificing optimality. Unfortunately, most error- or distance-based measures are not monotonic. A new measure is employed in this work that is monotonic and fast to compute. The search for relevant features according to this measure is guaranteed tobe complete but not exhaustive. Experiments are conducted for verification.