由数据构造贝叶斯网络结构是NP-难问题,根据互信息和条件独立测试,提出了一种构建最优贝叶斯网络结构的新算法。数值实验表明,新算法能较快地确定出与数据匹配程度最高的网络结构,从而能更高效地学习贝叶斯网络结构。
Constructing Bayesian network structures from data is NP-hard. According to the mutual information and conditional independence test, this paper presented a new algorithm for the construction of the optimal Bayesian network structure. Numerical experiments show that the new algorithm can determined much faster the structure with highest degree of data matching, thus the study of Bayesian network structures become more efficient.