An Objective Function for Belief Net Triangulation
Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics, PMLR R1:355-362, 1997.
This paper presents a new approach to the triangulation of belief networks. Triangulation is a combinatorial optimization problem; our idea is to embed its discrete domain into a continuous domain e. Then, by suitably extending the objective function over e, we can make use of continuous optimization techniques to do the minimization. We used an upper bound of the total junction tree weight as the cost function. The appropriateness of this choice is discussed and explored by simulations.