Constrained learning vector quantization or relaxed k-separability

Abstract

Neural networks and other sophisticated machine learning algorithms frequently miss simple solutions that can be discovered by a more constrained learning methods. Transition from a single neuron solving linearly separable problems, to multithreshold neuron solving k-separable problems, to neurons implementing prototypes solving q-separable problems, is investigated. Using Learning Vector Quantization (LVQ) approach this transition is presented as going from two prototypes defining a single hyperplane, to many co-linear prototypes defining parallel hyperplanes, to unconstrained prototypes defining Voronoi tessellation. For most datasets relaxing the co-linearity condition improves accuracy increasing complexity of the model, but for data with inherent logical structure LVQ algorithms with constraints significantly outperforms original LVQ and many other algorithms. © 2009 Springer Berlin Heidelberg.