A hybrid approach for learning parameters of probabilistic networks from incomplete databases

Authors:
Sajjad Haider
Affiliations:
System Architectures Lab, George Mason University, Fairfax, VA
Venue:
Design and application of hybrid intelligent systems
Year:
2003

Citing 7
Cited 2

Probabilistic reasoning in intelligent systems: networks of plausible inference

Probabilistic reasoning in intelligent systems: networks of plausible inference
Gibbs sampling in Bayesian networks (research note)

Artificial Intelligence
A Bayesian Method for the Induction of Probabilistic Networks from Data

Machine Learning
The EM algorithm for graphical association models with missing data

Computational Statistics & Data Analysis - Special issue dedicated to Toma´sˇ Havra´nek
Learning Bayesian Networks: The Combination of Knowledge and Statistical Data

Machine Learning
A Guide to the Literature on Learning Probabilistic Networks from Data

IEEE Transactions on Knowledge and Data Engineering
Learning Bayesian networks from incomplete databases

UAI'97 Proceedings of the Thirteenth conference on Uncertainty in artificial intelligence

Belief Functions Based Parameter and Structure Learning of Bayesian Networks in the Presence of Missing Data

International Journal of Hybrid Intelligent Systems
Theory of Influence Networks

Journal of Intelligent and Robotic Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Probabilistic Inference Networks are becoming increasingly popular for modeling and reasoning in uncertain domains. In the past few years, many efforts have been made in learning the parameters of probabilistic network directly from databases. The learning techniques are based on the assumption that the database is complete. If there are missing values, these values are first estimated and then the modified database is used to learn the parameters of the network. During the estimation process, the missing values are assumed to be missing at random. This paper incorporates the concepts of the Dempster-Shafer theory of belief functions to learn the parameter of the inference network. Instead of filling the missing values by their estimates, we model these missing values as representing our ignorance or lack of belief in their state. The representation provides the possibility to obtain interval estimates of the marginal probabilities of each variable in the inference network. The representation also allows us to model evidence in terms of support functions as used in belief functions, thus providing a richer way to enter evidence or new findings in the probabilistic model.