New specifics for a hierarchial estimator meta-algorithm

Abstract

Hierarchical Estimator is a meta-algorithm presented in [1] concerned with learning a nonlinear relation between two vector variables from training data, which is one of the core tasks of machine learning, primarily for the purpose of prediction. It arranges many simple function approximators into a tree-like structure in order to achieve a solution with a low error. This paper presents a new version of specifics for that meta-algorithm - a so called training set division and a competence function creation method. The included experimental results show improvement over the methods described in [1]. A short recollection of Hierarchical Estimator is also included. © 2012 Springer-Verlag Berlin Heidelberg.