DF-MAP:一种基于概率图模型的案件判决路径挖掘算法
|
摘要
通过获取案件的判决路径,法院判决系统可以轻松地对案件进行判决.然而,随着司法资源的迅猛增加以及案情特征的多样性,为快速获取案件判决路径提出了挑战.论文利用Rete算法在分析已有法律法规中可能存在的规则集合基础上,根据案件判决路径的有向性,提出了结合案情描述关键字和适用法律规则的概率图模型—Rete-PGM.根据Rete-PGM特征,利用有向图理论及最大后验概率查询算法,提出了适合于Rete-PGM特征的最有可能的路径挖掘算法—DF-MAP(Deep First Max A Posterior),并用实验验证了该算法的性能.通过将所提算法运用于真实的法律文书数据集,实现了真实案件的判决路径挖掘.该模型的提出以及案件判决路径的发现,为创建高效的法院判决系统提供了保障.
By accessing the case decision path,the court system can easily adjudicate the case. However,as the rapid increase of judicial resources and the diversity of case characteristics,it poses great challenges to obtain the case decision path. Based on the analysis of possible rules of in existing laws and the direction of the path,this work proposes a probability graph model based on Rete algorithm —Rete-PGM. According to the properties of Rete-PGM,using the theory of directed graph and max a posterior query algorithms,we present an algorithm —DF-MAP—for efficiently mining case decision path in Rete-PGM. This performance of the algorithm is verified by experiments. Through the application of the algorithm to real data sets,we successfully discover the case decision path. The model and the discovery of the case decision path provide a basis for quickly creating efficient judgment.
By accessing the case decision path,the court system can easily adjudicate the case. However,as the rapid increase of judicial resources and the diversity of case characteristics,it poses great challenges to obtain the case decision path. Based on the analysis of possible rules of in existing laws and the direction of the path,this work proposes a probability graph model based on Rete algorithm —Rete-PGM. According to the properties of Rete-PGM,using the theory of directed graph and max a posterior query algorithms,we present an algorithm —DF-MAP—for efficiently mining case decision path in Rete-PGM. This performance of the algorithm is verified by experiments. Through the application of the algorithm to real data sets,we successfully discover the case decision path. The model and the discovery of the case decision path provide a basis for quickly creating efficient judgment.
引文
[1]Wang P,Xu J,Xu B,et al.Semantic clustering and convolutional neural netw ork for short text categorization[J].Proceedings of the53rd Annual M eeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing,2015:352-357.
[2]Tharindu W,Nandula P,S.R.Liyanage.A method to extract essential keyw ords from a tw eet using NLP tools[C].International Conference on Advances in ICT for Emerging Regions,2016:029-034.
[3]Doorenbos R B.Production matching for large learning systems[J].Production Matching for Large Learning Systems,1995.
[4]Charles Forgy.A network match routine for production systems[R].Working Paper,1974.
[5]Charles Forgy.On the efficient implementation of production systems[D].Carnegie-M ellon University,1979.
[6]Charles,Forgy.Rete:a fast algorithm for the many pattern/many object pattern match problem[J].Artificial Intelligence,1982,19(1):17-37.
[7]Samuelsson C,Rayner M.Quantitative evaluation of explanationbased learning as an optimization tool for a large-scale natural language system[C].International Joint Conference on Artificial Intelligence(DBLP),1991:609-615.
[8]Liu Jian-wei,Li Hai-en,Luo Xiong-lin.Learning technique of probabilistic graphical models:a review[J].Acta Automatica Sinica,2014,40(6):1025-1044.
[9]Weiss Y,Freeman W T.On the optimality of solutions of the maxproduct belief propagation algorithm in arbitrary graphs[J].IEEE Transactions on Information Theory,2001,47(2):736-744.
[10]Rajaraman A,Ullman J D.Mining of massive datasets[M].Cambridge University Press,2012.
[11]Jagarlamudi J,Hal DauméI,Udupa R.Incorporating lexical priors into topic models[C].Conference of the European Chapter of the Association for Computational Linguistics,Association for Computational Linguistics,2012:204-213.
[12]Pawar D D,Bewoor M S,Patil S H.Text Rank:a novel concept for extraction based text summarization[J].International Journal of Computer Science and Information Technology,2014:3301-3304.
[13]Rose S,Engel D,Cramer N,et al.Automatic keyword extraction from individual documents[M].Text M ining:Applications and Theory.John Wiley&Sons,Ltd,2010:1-20.
[14]Deehter R Bucket elimination:a unifying framework for reasoning[J].Artificial Intelligence,1999,113(1/2):41-85.
[15]Lauritzen S L,Spiegelhalter D J.Local computations with probabilities on graphical structures and their application to expert systems[J].Journal of the Royal Statistical Society,1988:157-224.
[8]刘建伟,黎海恩,罗雄麟,等.概率图模型学习技术研究进展[J].自动化学报,2014,40(6):1025-1044.
1China Judgements Online. http://wenshu. court. gov. cn,2017.
2 The Supreme People's Court of the People's Republic of China. http://www. court. gov. cn/fabu-xiangqing-18362. html,2015.
3http://hanlp. linrunsoft. com/
[2]Tharindu W,Nandula P,S.R.Liyanage.A method to extract essential keyw ords from a tw eet using NLP tools[C].International Conference on Advances in ICT for Emerging Regions,2016:029-034.
[3]Doorenbos R B.Production matching for large learning systems[J].Production Matching for Large Learning Systems,1995.
[4]Charles Forgy.A network match routine for production systems[R].Working Paper,1974.
[5]Charles Forgy.On the efficient implementation of production systems[D].Carnegie-M ellon University,1979.
[6]Charles,Forgy.Rete:a fast algorithm for the many pattern/many object pattern match problem[J].Artificial Intelligence,1982,19(1):17-37.
[7]Samuelsson C,Rayner M.Quantitative evaluation of explanationbased learning as an optimization tool for a large-scale natural language system[C].International Joint Conference on Artificial Intelligence(DBLP),1991:609-615.
[8]Liu Jian-wei,Li Hai-en,Luo Xiong-lin.Learning technique of probabilistic graphical models:a review[J].Acta Automatica Sinica,2014,40(6):1025-1044.
[9]Weiss Y,Freeman W T.On the optimality of solutions of the maxproduct belief propagation algorithm in arbitrary graphs[J].IEEE Transactions on Information Theory,2001,47(2):736-744.
[10]Rajaraman A,Ullman J D.Mining of massive datasets[M].Cambridge University Press,2012.
[11]Jagarlamudi J,Hal DauméI,Udupa R.Incorporating lexical priors into topic models[C].Conference of the European Chapter of the Association for Computational Linguistics,Association for Computational Linguistics,2012:204-213.
[12]Pawar D D,Bewoor M S,Patil S H.Text Rank:a novel concept for extraction based text summarization[J].International Journal of Computer Science and Information Technology,2014:3301-3304.
[13]Rose S,Engel D,Cramer N,et al.Automatic keyword extraction from individual documents[M].Text M ining:Applications and Theory.John Wiley&Sons,Ltd,2010:1-20.
[14]Deehter R Bucket elimination:a unifying framework for reasoning[J].Artificial Intelligence,1999,113(1/2):41-85.
[15]Lauritzen S L,Spiegelhalter D J.Local computations with probabilities on graphical structures and their application to expert systems[J].Journal of the Royal Statistical Society,1988:157-224.
[8]刘建伟,黎海恩,罗雄麟,等.概率图模型学习技术研究进展[J].自动化学报,2014,40(6):1025-1044.
1China Judgements Online. http://wenshu. court. gov. cn,2017.
2 The Supreme People's Court of the People's Republic of China. http://www. court. gov. cn/fabu-xiangqing-18362. html,2015.
3http://hanlp. linrunsoft. com/