Markov逻辑网在链接预测中的应用  

Markov logic networks based link prediction

作　　者：张玉芳[1] 孔润[1] 熊忠阳[1] 田源[1] 舒方俊[1] 

机构地区：[1]重庆大学计算机学院,重庆400044

出　　处：《计算机应用研究》2011年第6期2154-2157,共4页Application Research of Computers

基　　金：重庆市自然科学基金资助项目(CSTC2008BB2191)

摘　　要：传统同类独立同概率分布的链接预测方法会带来很大的噪声,导致预测效果很差,将Markov逻辑网应用到链接预测中,旨在改善这一问题。Markov逻辑网是将Markov网与一阶逻辑结合的统计关系学习方法。利用Markov逻辑网构建关系模型,确定实体之间是否存在链接关系以及当链接关系存在时预测此链接关系的类型。针对两个数据集的实验结果,显示了采用Markov逻辑网模型要比传统链接预测模型有更好的效果,进而为Markov逻辑网解决实际问题提供了依据。The traditional identically,independent and identically distributed（IID） model brings a lot of noise,resulting in bad performance in link prediction.In order to solve this problem,this paper used Markov logic networks（MLNs） in link prediction.Markov logic networks was a new statistical relational learning（SRL） model in which Markov networks and the first-order logic combined together.It focused on applying Markov logic networks to build a relational model,so as to predict the existence and the type of links between entities.Experiments results based on applying Markov logic networks model to two datasets prove it has better performance than traditional link prediction models,which providing evidences for Markov logic networks solving practical problems as well.

关 键 词：链接预测 MARKOV逻辑网 MARKOV网 统计关系学习 机器学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]