详细信息
文献类型:期刊文献
中文题名:Bayesian网转化为神经元网
英文题名:Compiling Bayesian Networks into Neural Networks
作者:薛万欣[1];董冠宇[2];刘大有[3]
第一作者:薛万欣
通讯作者:Xue, W.-X.
机构:[1]北京联合大学管理学院;[2]北京后勤指挥学院;[3]吉林大学计算机科学技术学院
第一机构:北京联合大学管理学院
通讯机构:[1]Coll. of Mgmt. Eng., Beijing Union Univ., Beijing 100101, China|[11417]北京联合大学;
年份:2004
卷号:32
期号:2
起止页码:250-253
中文期刊名:电子学报
外文期刊名:Acta Electronica Sinica
收录:CSTPCD;;EI(收录号:2004268238224);Scopus(收录号:2-s2.0-2942674764);北大核心:【北大核心2000】;CSCD:【CSCD2011_2012】;
基金:国家 8 63项目 (No.863 30 6 ZD0 5 0 1 2 ) ;国家自然科学基金 (No.698830 0 3) ;教育部高校博士点专项科研基金项目;教育部符号计算与知识工程重点实验室项目
语种:中文
中文关键词:Bayesian网;SIGMA-PI网络;后向传播网
外文关键词:Mathematical models;Neural networks;Probability distributions
摘要:Bayesian网目前广泛应用于专家系统中 ,用于处理大量以条件概率为形式的数据 .本文借用神经元网络结构 ,根据专家给定的相关模型和部分观察集使用后向传播对条件概率进行估计 ,并在训练中 ,保持Bayesian网特性不变 ,应用Occam修剪法则 ,在化简过程中提炼其中的规律 .实践表明 ,对于复杂的问题 。
The criticism on the usage of Bayesian networks in expert systems is centered around the claim that the use of probability requires a massive amount of data in the form of conditional probabilities. This paper shows that with given information easily obtained from experts, the dependence model probabilities can be estimated using backpropagation, such that during training the Bayesian characteristic of the network is preserved. Applying the Occam's razor principle results in defining a partial order among neural network structures. Experiments show that for the Multiplexer problem, the network compiled from the more succinct causal model is better than the one compiled from the less succinct model.
参考文献:
正在载入数据...