Bayesian Assaying of GUHA Nuggets

Abstract

The General Unary Hypothesis Automaton (GUHA) sifts through large data sets and finds 2×2 contingency tables satisfying certain dependencies expressible by generalized quantifiers that indicate possible relations between attributes. In this work we show how these tables can be further investigated by Bayesian statistical methods. In this way we are able to translate tables that have been discovered by GUHA into verbal statements and probability density plots that are comprehensible to anyone who has a basic understanding of probability. © Springer-Verlag Berlin Heidelberg 2010.