Learning Bayesian Networks from Incomplete Data with Stochastic Search\n Algorithms

2013/01/23 by James W. Myers, Myers, James W., Kathryn Blackmond Laskey +3
Computer Science · Mathematics · #Artificial Intelligence (cs.AI) #Bayesian Methods and Mixture Models #Bayesian Modeling and Causal Inference #FOS: Computer and information sciences #Machine Learning (cs.LG) #Neural Networks and Applications #Statistical Methods and Bayesian Inference #cs.AI #cs.LG

paper · pdf · doi:10.48550/arxiv.1301.6726

Appears in Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence (UAI1999)

arxiv created 2013/01/23 · openalex publication_date 2013/01/23 · arxiv updated 2013/01/30 · openalex created_date 2025/10/24 · openalex updated_date 2026/07/28

Abstract

This paper describes stochastic search approaches, including a new stochastic\nalgorithm and an adaptive mutation operator, for learning Bayesian networks\nfrom incomplete data. This problem is characterized by a huge solution space\nwith a highly multimodal landscape. State-of-the-art approaches all involve\nusing deterministic approaches such as the expectation-maximization algorithm.\nThese approaches are guaranteed to find local maxima, but do not explore the\nlandscape for other modes. Our approach evolves structure and the missing data.\nWe compare our stochastic algorithms and show they all produce accurate\nresults.\n

Citations

• The Bayesian Structural EM Algorithm
• Equation of State Calculations by Fast Computing Machines
• Monte Carlo sampling methods using Markov chains and their applications
• Inference from Iterative Simulation Using Multiple Sequences

Related

• Learning Bayesian Networks with Incomplete Data by Augmentation
• Efficient Algorithms for Bayesian Network Parameter Learning from Incomplete Data
• Improving parameter learning of Bayesian nets from incomplete data
• Learning Bayesian Network Structure from Massive Datasets: The "Sparse\n Candidate" Algorithm
• The AI&M Procedure for Learning from Incomplete Data