Logic sampling, likelihood weighting and AIS-BN : an exploration of importance sampling

Graduation date: 2002Logic Sampling, Likelihood Weighting and AIS-BN are three variants of\ud stochastic sampling, one class of approximate inference for Bayesian networks.\ud We summarize the ideas underlying each algorithm and the relationship among\ud them. The results from a set of empirical experiments comparing Logic Sampling,\ud Likelihood Weighting and AIS-BN are presented. We also test the impact\ud of each of the proposed heuristics and learning method separately and in combination\ud in order to give a deeper look into AIS-BN, and see how the heuristics\ud and learning method contribute to the power of the algorithm.\ud Key words: belief network, probability inference, Logic Sampling, Likelihood\ud Weighting, Importance Sampling, Adaptive Importance Sampling Algorithm for\ud Evidential Reasoning in Large Bayesian Networks(AIS-BN), Mean Percentage\ud Error (MPE), Mean Square Error (MSE), Convergence Rate, heuristic, learning\ud method