基于自然语义处理的裁判文书分割系统  

Referee document segmentation system based on natural semantic processing

作　　者：郑少婉 陆培民[1] 

机构地区：[1]福州大学物理与信息工程学院,福建福州350002

出　　处：《信息技术与网络安全》2018年第2期55-57,共3页Information Technology and Network Security

摘　　要：研究了基于语义的裁判文书成分分割的方法,旨在在语义理解的基础上,从裁判文书中获取判决书信息、原告信息、被告信息、案件事实、原告诉求、被告辩称、法院认定证据、法院观点、法律依据、判决结果等信息,从而减少法官工作量,以及帮助人们更好地了解案件。通过两个部分详解了系统的构建以及分割的实现,系统构建主要包括线下裁判文书采集、语料库训练,以及线上的分析模块。This paper studies a method of judgment component of natural semantic processing based on segmentation. Based on semantic understanding it obtains information, including the plaintiff and defendant, defendant information, the facts of the case, plaintiff appeals, defendant argued, the verdict and other information from the judgment, thus reducing the workload of judges, and helping people to better understand the cases. The construction of the system and the implementation of the segmentation are explained in two parts. The system construction includes the following modules： the collection of referee documents, corpus training and online analysis module.

关 键 词：Word2vec 中文分词 文书分割 相似度计算 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]