對於法律問題進行判例檢索和法條預測;Judgment Retrieval and Statute Prediction for Legal Problems

NCU Institutional Repository > 管理學院 > 資訊管理研究所 > 博碩士論文 > Item 987654321/66887

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/66887

题名:	對於法律問題進行判例檢索和法條預測;Judgment Retrieval and Statute Prediction for Legal Problems
作者:	劉譯閎;Liu,Yi-Hung
贡献者:	資訊管理學系
关键词:	文件探勘;法條;刑事判例;向量空間模型;標準化谷歌距離;支援向量機;Text Mining;Statute;Criminal judgment;Vector space model;Normalized Google Distance;Support Vector Machines
日期:	2014-11-24
上传时间:	2015-03-16 15:58:14 (UTC+8)
出版者:	國立中央大學
摘要:	應用文件探勘在法律問題處理上已成為近年來新興的研究領域。就我們所知，即使先前已有少數的研究著重於協助法律專業人士檢索相關的法律文件，然而這些研究並未考量到一般人民在使用法律詞彙來描述碰到的法律問題有其困難的地方，同時，也沒有研究在探討有關於利用法律問題來進行相關法條預測。在本論文中，我們探討二個研究議題：藉由運用法律文件的特性進行判例檢索及法條預測。在第一個研究主題之中，我們提出了一個基於文件探勘的方法讓一般人士可以使用日常詞彙來搜尋及檢索出相關的刑事判例。在第二個研究主題中，提出了一個三階段法條預測方法。這個預測的方法提供非專業人士使用日常詞彙來描述法律問題進而用以找出問題所涉及相關的法條。本文透過兩個主要實驗設計來驗證成效。在第一個研究議題實驗上，我們使用了傳統的TF-IDF方法與本文所提出的判例檢索方法透過問卷調查的方式進行成效比較。就第二個研究議題實驗中，我們採用了四個知名的檢索方法分別為Cosine 相似度、Pearson 相關係數、 Spearman′s相關係數及TF-IDF與本文提出的三階段法條預測方法進行成效比較。經由實驗過程(以中文刑事判例為資料集)，說明這兩個研究議題所提出的方法皆具有效性及準確性，同時顯示此兩個方法皆優於傳統方法。;Applying text mining techniques to legal issues has been an emerging research topic in recent years. Although a few previous studies focused on assisting professionals in the retrieval of related legal documents, to our knowledge, they did not take into account the general public and their difficulty in describing legal problems in professional legal terms and could not provide relevant statutes to the general public using problem statements. In this dissertation, we formulate two research topics: judgment retrieval and statute prediction using the unique characteristics of legal documents. In the first research topic, we design a text mining based method that allows the general public to use everyday vocabulary to search for and retrieve criminal judgments. Then we present an innovative approach, the three-phase prediction (TPP) algorithm, which enables laypeople to use daily vocabulary to describe their problems and find pertinent statutes for their cases. There are two experiments to validate our proposed research methods. The first experimental study compares the performances of traditional TF-IDF method and our judgment retrieval approach through a survey. The second one is based on the statute prediction problem, and four state of the art retrieval functions including Cosine similarity, Pearson correlation coefficient, Spearman′s correlation coefficient and TF-IDF methods are compared with TPP. Both proposed methods have been verified for accuracy and effectiveness by using Chinese Criminal Code judgments. The results show that the proposed methods are accurate and they are more advantageous than traditional methods.
显示于类别:	[資訊管理研究所] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	745	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....