Evolutionary Methods for Learning Bayesian Network Structures
Thierry Brouard, Alain Delaplace, Hubert Cardot · InTech eBooks · 2008
operator which ensures the closeness of the previous.Various strategies were then tested in order to find a balance between speed of convergence and avoidance of local optima.We focus particularly onto two of these: a new adaptive scheme to the mutation rate on one hand and sequential niching techniques on the other.The remaining of the chapter is structured as follows: In the second section we will define the problem, ended by a brief state of the art.In the third section, we will show how an evolutionary approach is well suited to this kind of problem.After briefly recalling the theory of genetic algorithms, we will describe the representation of a Bayesian network adapted to genetic algorithms and all the needed operators necessary to take in account the inherent constraints to Bayesian networks.In the fourth section the various strategies will then be developed: Adaptive scheme to the mutation rate on one hand and niching techniques on the other hand.The fifth section will describe the test protocol and the results obtained compared to other classical algorithms.A study of the behaviour of the used strategies will also be given.And finally, the sixth section will present future search in this domain.