This article is cited in 13 papers

Minimal join graph set synthesis self-managing cliques algorithm

A. A. Fil'chenkov<sup>ab

a</sup> St. Petersburg State University, Department of Mathematics and Mechanics
^b St. Petersburg Institute for Informatics and Automation of RAS

Abstract: The scheme of an algorithm allowing to design minimal join graph set from a given maximal knowledge pattern set is known, but it can be improved by engaging the developed theory of algebraic Bayesian network global structure. The goal of the work is to improve the known minimal join graph set algorithm. Three improvements of the known algorithm are suggested and justified: exclusion of not significant narrowing, exclusion of cliques with the only possession and a prior childless cliques with the only edge processing. A new algorithm implementing suggested improvements is designed.

Keywords: algebraic Bayesian networks, secondary structure, machine learning, proba-bilistic graphical models.

UDC: 004.8

Received: 06.12.2010
Accepted: 06.12.2010