Structure Learning of Bayesian Network with Latent Variables by Weight-Induced Refinement

Abstract

Bayesian network (BN) with latent variables (LVs) provides a concise and straightforward framework for representing and inferring uncertain knowledge with unobservable variables or with regard to missing data. To learn the BN with LVs consistently with the realistic situations, we propose the information theory based concept of existence weight and incorporate it into the clique-based learning method. In line with the challenges when learning BN with LVs, we focus on determining the number of LVs, and determining the relationships between LVs and the observed variables. First, we define the existence weight and propose the algorithms for finding the ε-cliques from the BN without LVs learned from data. Then, we introduce the LV to each ε-clique and adjust the BN structure with LVs. Further, we adjust the value of parameter ε to determine the number of LVs. Experimental results show the feasibility of our method.

Hidden Variable Discovery Based on Regression and Entropy

“…We defined a causal intensity indicator and found that the causal intensity of two points with direct causality is significantly greater than the causal intensity of two points that are jointly affected by the hidden variables and do not have direct causality. The causal intensity [19] is denoted as:…”

Web-KR 2014

“…In [4], the authors investigate on ontology matching based on learning using partially labeled data, while [5] presents a structure learning algorithm for uncertain knowledge representation and inference.…”