摘要
传统的贝叶斯网络结构学习是基于完备数据集以及数据量较小的学习,随着大数据以及云计算的发展,呈现出数据量大且复杂的特点,但是在不完备数据集上的贝叶斯网络结构学习是不足的。而IMS-EM算法则分别从数据缺失情况的不完备数据集以及存在隐藏变量的不完备数据集两个方面进行实验验证,实验结果显示:该算法在不完备数据集上的贝叶斯网络结构学习具有较好的优势,对于不完备数据集上的结构学习具有一定的可行性。
Traditional structure learning of Bayesian networks is based on complete data sets and small amount of data.Along with the development of big data and cloud computing,it presents the characteristics of large and complex data.However,structure learning of Bayesian networks on incomplete data sets is insufficient.The IMS-EM algorithm is experimentally verified from two aspects:incomplete data set with missing data and incomplete data set with hidden variables.The experimental results show that the algorithm has a good advantage in structure learning of Bayesian networks on incomplete data sets with certain feasibility.
作者
李昌群
李现伟
蔡政策
LI Changqun;LI Xianwei;CAI Zhengce(Information Technology Department,Anhui Vocational College of Grain Engineering,Hefei 230000,China;School of Computer Science,Wuhan University,Wuhan 430072,China;School of Computer Science and Information Engineering,Bengbu University,Bengbu 233000,China;Information Engineering Department,Anhui Institute of International Business,Hefei 230000,China)
出处
《浙江水利水电学院学报》
2023年第1期81-85,共5页
Journal of Zhejiang University of Water Resources and Electric Power
基金
2022年高校优秀青年骨干教师国内访问研修项目(GXGNFX2022213)
安徽省自然科学基金重点资助项目(KJ2021A1562)
安徽省质量工程重点项目(2021JXTD088)
安徽粮食工程职业学院精品课程(AHLY2022008)。
关键词
贝叶斯网络
不完备数据集
结构学习
Bayesian network
incomplete data set
structure learning
