Markov逻辑网研究综述  

Survey of Markov Logic Networks

作　　者：徐元子[1] 张迎新[2] 刘登第[1] 

机构地区：[1]空军指挥学院,北京 [2]中国人民解放军71375部队,山东潍坊

出　　处：《软件工程与应用》2015年第3期73-80,共8页Software Engineering and Applications

摘　　要：Markov逻辑网是将Markov网络与一阶谓词逻辑相结合的统计关系学习模型。Markov逻辑网在实体识别、数据融合、信息抽取等领域都有重要研究价值,具有广泛的应用。本文较为全面的介绍了Markov逻辑网的理论模型、推理、参数学习、与其他算法的比较,最后探讨Markov逻辑网未来的研究方向。Markov logic networks (MLNs) is a kind of statistical relational learning model which combines Markov network and first-order logic together. MLNs has the significant research value and in many areas it has widely applications, such as entity recognition, data integration and information extrac-tion. In this paper, we introduced the theoretical model of Markov logic networks, inference and pa-rametric learning of it and compared it with other. In the end, we discussed future works of MLNs.

关 键 词：MARKOV逻辑网 一阶谓词逻辑 统计关系学习 推理 参数学习 

分 类 号：TP1[自动化与计算机技术—控制理论与控制工程]