Learning Bayesian Network Structure using Markov Blanket in K2 Algorithm

ýA Bayesian network is a graphical model that represents a set of random variables and their causal relationship via a Directed Acyclic Graph (DAG)ý. ýThere are basically two methods used for learning Bayesian networký: ýparameter-learning and structure-learningý. ýOne of the most effective structure-learning methods is K2 algorithmý. ýBecause the performance of the K2 algorithm depends on node orderingý, ýmore effective node ordering inference methods are neededý. ýIn this paperý, ýbased on the fact that the parent and child variables are identified by estimated Markov Blanket (MB)ý, ýwe first estimate the MB of a variable using Grow-Shrink algorithmý, ýthen determine the candidate parents of a variable by evaluating the conditional frequencies using Dirichlet probability density functioný. ýThen the candidate parents are used as input for the K2 algorithmý. ýExperimental results for most of the datasets indicate that our proposed method significantly outperforms previous methodý.