计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 220800046-7.doi: 10.11896/jsjkx.220800046

• 大数据&数据科学 • 上一篇    下一篇

贝叶斯网络结构学习的CMIHC算法

李晓晴, 于海征   

  1. 新疆大学数学与系统科学学院 乌鲁木齐 830046
  • 发布日期:2023-11-09
  • 通讯作者: 于海征(yuhaizheng@xju.edu.cn)
  • 作者简介:(812915644@qq.com)
  • 基金资助:
    国家自然科学基金(61662079,11761070,U1703262);新疆自治区自然科学基金面上项目(2021D01C078)

CMIHC Algorithm for Bayesian Network Structure Learning

LI Xiaoqing, YU Haizheng   

  1. College of Mathematics and Systems Science,Xinjiang University,Urumqi 830046,China
  • Published:2023-11-09
  • About author:LI Xiaoqing,born in 1996,postgra-duate.Her main research interests include Bayesian network structure learning and so on.
    YU Haizheng,born in 1976,Ph.D,associate professor,is a member of China Computer Federation.His main research interests include big data analysis and so on.
  • Supported by:
    National Natural Science Foundation of China(61662079,11761070,U1703262) and Xinjiang Natural Science Foundation(2021D01C078).

摘要: 贝叶斯网络源于对人工智能领域不确定问题的研究,是进行不确定问题推理和数据分析的重要工具。自贝叶斯网络结构学习诞生以来,已有众多成熟的结构学习算法,包括基于依赖分析的方法、基于评分搜索的方法和混合搜索的方法。其中利用信息论进行结构修剪已成为常用手段,但条件互信息中条件集的选取并没有统一标准,导致网络结构的修剪不一致。爬山算法利用3种搜索算子对网络结构进行局部更新,通过评分函数得到最优结构。结合信息论和爬山算法思想,提出一种新的结构学习算法——CMIHC(Conditional Mutual Information Hill Climbing)算法。该算法利用互信息和创建的条件集修剪初始连通图,对其进行定向,进而得到初始网络结构,结合评分函数和爬山算法的贪婪搜索策略得到最优网络结构。通过实验分析,在精度和效率上,CMIHC算法效果优于其他对比算法。

关键词: 贝叶斯网络, 结构学习, 条件互信息, 爬山算法, CMIHC算法

Abstract: Bayesian network originates from the research on uncertain problems in the field of artificial intelligence.It is an important tool for reasoning and data analysis of uncertain problems.Since the birth of Bayesian network structure learning,there have been many mature structure learning algorithms,including dependency analysis based method,score based search method and hybrid search method.Among them,structure pruning by information theory has become a common method,but there is no unified standard for the selection of condition set in conditional mutual information,resulting in inconsistent pruning of network structure.The hill climbing algorithm uses three search operators to update the network structure locally,and obtains the optimal structure through the scoring function.Combined with the idea of information theory and hill climbing algorithm,a new structure learning algorithm-conditional mutual information hill climbing(CMIHC) algorithm is proposed.The proposed algorithm prunes the initial connected graph by using mutual information and the created condition set,and orients it to obtain the initial network structure.Combined with the scoring function and the greedy search strategy of hill climbing algorithm,the optimal network structure is obtained.Experimental analysis shows that CMIHC algorithm is superior to other comparison algorithms in accuracy and efficiency.

Key words: Bayesian network, Structure learning, Conditional mutual information, Hill climbing algorithm, CMIHC algorithm

中图分类号: 

  • TP181
[1]ZHANG L W,GUO H P.Introduction to Bayesian Networks[M].Beijing:Science Publishing House,2006.
[2]LOU C Y,LI X S,ATOUI M A.Bayesian network based on an adaptive threshold scheme for fault detection and classification[J].Industrial and Engineering Chemistry Research,2020,59(34):15155-15164.
[3]TAGHI-MOLLA A,RABBANI M,KARIMI GAVARESHKIM H,et al.Safety improvement in a gas refinery based on resilience engineering and macro-ergonomics indicators:a Bayesian network-artificial neural network approach[J].Intermational Journal of System Assurance Engineering and Management,2020,11(3):641-654.
[4]WANG H,GUO X D,ZHU H.Research on causal reasoning method of domain ontology based on Bayesian network[J].Application Research of Computers,2019,36(3):711-715.
[5]LU E S,HAN M G,LIU F,et al.Analysis of Chinese medicine syndrome rules of severe pneumonia based on Bayesian network[J].Acta Chinese Medicine,2022,37(1):173-179.
[6]CHENG L X,YAN J N,JIAO Y,et al.Bayesian network methodfor remote sensing cloud user behavior authentication[J].Application Research of Computers,2019,36(2):441-445.
[7]LI S H,ZHANG J.Review of Bayesian networks structurelearning[J].Application Research of Computers,2015,32(3):641-646.
[8]SPIRTES P,GLYMOUR C,SCHEINES R.Causality fromProbability[M].London:Pima,Evolving Knowledge in Natural Science and Artificial Intelligence,1990:181-199.
[9]SPIRTES P,GLYMOUR C.An algorithm for fast recovery of sparse causal graphs[J].Social Science Computer Review,1991,9(1):62-72.
[10]CHENG J,GREINER R,KELLY J,et al.Learning Bayesiannetworks from data:An information-theory based approach[J].Artificial Intelligence,2002,137(1/2):43-90.
[11]XU P F,YANG Z.Accelerating GDS algorithm with L1 penalty for Bayesian network[J].Journal of Changchun University of Technology,2020,41(2):118-121.
[12]XU P F,WANG S D,SHANG L X,et al.Structural learning of Gaussian Bayesian network by bootstrap[J].Journal of Changchun University of Technology,2020,41(4):313-321.
[13]ZHAO J Z,WU C N,WANG X W,et al.Structure Learning Algorithm of Bayesian Networks Based on Markov Blanket[J].Journal of Northeastern University(Natural Science),2020,41(4):464-469,481.
[14]HERSKOVITS E.Computer-based probabilistic-network con-struction[D].Stanford University,1991.
[15]HECKERMAN D,GEIGER D,CHICKERING D M.Learning Bayesian networks:The combination of knowledge and statistical data[J].Machine learning,1995,20(3):197-243.
[16]GUO Z,CONSTANTINOU A C.Approximate learning of high dimensional Bayesian network structures via pruning of candidate parent sets[J].Entropy,2020,22(10):1142.
[17]BEHJATI S,BEIGY H.Improved K2 algorithm for Bayesian network structure learning[J].Engineering Applications of Artificial Intelligence,2020,91:103617.
[18]LV Y,MIAO J,LIANG J,et al.BIC-based node order learning for improving Bayesian network structure learning[J].Frontiers of Computer Science,2021,15(6):1-14.
[19]TSAMARDINOS I,BROWN E L,ALIFERIS C F.The max-min hill-climbing Bayesian network structure learning algorithm[J].Machine Learning,2006,65(1):31-78.
[20]LIU X,GAO X,WANG Z,et al.Improved local search with momentum for Bayesian networks structure learning[J].Entropy,2021,23(6):750.
[21]SUN B,ZHOU Y,WANG J,et al.A new PC-PSO algorithm for Bayesian network structure learning with structure priors[J].Expert Systems With Applications,2021,184:115237.
[22]WANG Y,TAN S Q,LIU Y H.Bayesian network structural learning algorithm based on mutual information[J].Computer Engineering,2011,37(7):62-64.
[23]LI B H,GAO X L,LIU S Y,et al.Learning Bayesian network structures based on mutual information[J].CAAI Transactions on Intelligent Systems,2011,6(1):68-72.
[24]JIN Y,HU Y A,ZHANG J,et al.Bayesian network structurelearning combining mutual information with hill climbing algorithm[J].Computer Applications and Software,2012,29(9):122-125.
[25]WU Y G,PANG S C.K2&HC Structure learning algorithm[J].Computer and Digital Engineering,2014,42(7):1137-1140,1145.
[26]LIU H R,LI X,MA M,et al.Simplified Greedy Algorithm for Bayesian Network Structure Learning[J].Journal of Chinese Computer Systems,2015,36(2):306-309.
[27]LIU B,LIU Y J,LIU H R,et al.A study of fault diagnosis mo-del of rotary kiln based on improved structural algorithm of Bayesian[J].China Mechanical Engineering,2017,28(18):2143-2151.
[28]LONG Y G.Learning Bayesian network structure basing on Information theory and causal effect[D].Changchun:Jilin University,2020
[29]GAO X G,WANG C F,DI R H.A block learning algorithmwith improved K-means algorithm for learning sparse BN optimal structure[J].Acta Automatica Sinica,2020,46(5):923-933.
[30]WU B W.Hybrid uncertainty Bayesian networks learning model and implementation by R[D].Guangzhou:South China University of Technology,2013.
[31]ZHAO J Z,WU C N,WANG X W,et al.Structure learning algorithm of Bayesian networks based on markov blanket[J].Journal of Northeastern University(Natural Science),2020,41(4):464-469,481.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!