Research On Clump Tree Reasoning Of Bayesian Network

In recent years,probabilistic graph models are widely used in many fields such as medical diagnosis,fault diagnosis,gene and chromosome data analysis,communication and coding.In the probability graph model,the clique tree structure is an active research field.We have studied the reasoning based on clique trees.On the basis of clique trees,we deeply discuss the algorithm of clique trees.In high dimensional data sets,there is generally insufficient data from which to characterize probabilities;the available data points are spread too thinly over a very large space of possible attribute combinations.However,the clique tree expresses dependencies in a high dimensional attribute space and can be used to make probabilistic inferences about data.In order to construct a clique tree,we need to find a maximal clique in the network.We use a maximal clique search algorithm to determine all the maximal cliques in a network,and then construct a clique tree structure through the maximal clique,a joint probability distribution in a high dimensional space can be decomposed into a product of lower dimensional probabilities.With in low dimensional spaces,the data is more concentrated and a probability distribution can be successfully derived.