進階搜尋


下載電子全文  
系統識別號 U0026-1607201901241000
論文名稱(中文) 結合案件資料與文本訊息預測調解案件之成功與否
論文名稱(英文) Predicting the Success of Mediation Requests Based on Heterogeneous Case Properties and Textual Information
校院名稱 成功大學
系所名稱(中) 電腦與通信工程研究所
系所名稱(英) Institute of Computer & Communication
學年度 107
學期 2
出版年 108
研究生(中文) 黃新幃
研究生(英文) Xin-Wei Huang
學號 Q36051059
學位類別 碩士
語文別 英文
論文頁數 31頁
口試委員 指導教授-解巽評
口試委員-黃仁暐
口試委員-李政德
口試委員-李嘉岩
中文關鍵字 法律預測  機器學習  特徵工程  自然語言處理 
英文關鍵字 Legal case predict  Machine learning  Feature engineering  Natural language processing 
學科別分類
中文摘要 在日常生活當中,常常有各式各樣的糾紛發生,當糾紛發生時人們往往傾向於去尋求第三方公正人士或機構來幫忙釐清糾紛,調解委員會便是其中一個管道,調解委員會的主要目的是有效解決人民之間的糾紛,從而減輕法院的負擔。但是調解過程往往會耗費大量時間和人力最後卻以失敗告終,如此一來民眾還是得回歸到一開始的情況去尋求正式的法律援助。另一方面,調解案件最後的成功與否受到諸多因素的影響,例如爭吵雙方的背景、人格以及調解員的談判技巧,這導致了種種不確定性,因此難以直觀的判定。本文採用與以往法律預測任務不同的方法,我們分析並預測調解委員會當中的爭議案件是否會在不久的將來得到解決,也就是說雙方透過調解委員會的調解而達成和平的協議。準確的預測調解結果的好處是雙重的,對於民眾而言,可以透過預測結果來考慮是否尋求調解,避免浪費時間,再者我們的模型還能更進一步地選出最有可能成功解決爭議的調解員。
關於現有的法律預測任務主要專注在刑事案件上,與以往研究的不同之處在於我們的案件主要都是民事案件,在這項工作當中,我們結合從調解委員會紀錄的應用程式中提取案件的相關信息和文本描述來預測調解結果。此外對於文本描述我們應用各種最新的文本探勘技術並加以比較。我們的實驗表明,結合這兩種信息的預測結果在F-measure的評估下可以達到84%的效果。此外我們更進一步擴展我們的預測模型,使其實現一個智能功能,可以根據用戶的調解請求推薦適當的調解員。這種功能能夠幫助調解委員會選擇最適當的調解員,而非如傳統隨機選派。實驗結果表明我們的推薦系統在測試資料當中可以成功的判斷95%的調解案例。總體而言,我們實施了一個綜合系統,能預測調解案件請求最終是否會成功,此外更推薦適當的調解委員。
英文摘要 The main purpose of the mediation committee is to effectively resolve the disputes between people so that the burden of the court can be relieved. However, a mediation process sometimes consumes much time and human effort. The worst result of a mediation is failure. If such a case happens, people still need to seek formal legal assistance. On the other hand, the success of mediation is affected by many factors, such as the context of the quarrel, personality of both parties, and the negotiation skill of the mediator, which leads to uncertainty, thus it can hardly be predicted directly. This paper takes a different approach from the previous legal prediction research. It analyzes and predicts whether a dispute case in the mediation committee will be resolved in the near future, which means two parties reach an agreement peacefully through the conciliation of the mediator. The benefits of predicting the success of mediation are two-fold. First, the parties in a case can consider whether to seek for mediation based on the inference result, which can avoid wasting time if the model has a negative prediction. Second, such inference can further help us to select the leading mediator who is most likely to successfully resolve the dispute.
Existing works about legal case prediction mostly focused on prosecution or criminal cases. In this work, we combine the case information and textual features extracted from mediation applications to predict the result of mediation. In addition, for textual features, we apply different state-of-art text mining models and show the comparisons. Our experiments show that combining such two kinds of information can effectively achieve 84% for F-measure. In addition, we further extend the inference model to implement an intelligent function that can recommend suitable mediators based on users’ mediation requests. Such function is quite useful for committees to select a right mediator for different cases. Our experiments verified our recommender system could judge successfully 95% of mediation cases in testing data. We implemented a system for joint functions of predicting the success of mediation request and recommending mediators for Tainan City Government to help mediate disputes.
論文目次 Table of Contents
Abstract V
List of Tables X
List of Figures XI
Chapter 1 Introduction 1
Chapter 2 Related Work 4
Chapter 3 Datasets Construction 5
Chapter 4 Feature Engineering 7
4.1 Case Information (CI) 7
4.2 Text Information 9
Chapter 5 MEDIATOR RECOMMENDATION 13
5.1 FEATURE EXTENSION FOR MEDIATORS 13
5.2 RECOMMENDATION 15
Chapter 6 Experiments 16
6.1 Baseline 16
6.2 Implementation Details 17
6.3 Metrics 17
6.4 Experiments of predict 18
6.5 Experiments of text vector dimension on predict task 22
6.6 Experiments of Recommended mediator 25
6.7 Experiments of text dimension on Recommended mediator 27
Chapter 7 Conclusion 29
References 30

參考文獻 [1] S. Arora, Y. Liang, T. Ma. A simple but tough-to-beat baseline for sentence embedding. ICLR 2017.
[2] N. Aletras, D. Tsarapatsanis, D. Preoţiuc-Pietro and V. Lampos. 2016. Predicting judicial decisions of the european court of human rights: A natural language processing perspective. PeerJ Computer Science 2.
[3] M. Elhoseiny, B.Saleh, and A. Elgammal. 2014. Write a classifier: Zero-shot learning using purely textual descriptions. In Proceedings of ICCV2014.
[4] Kort F. 1957. Predicting Supreme Court decisions mathematically: a quantitative analysis of the ‘‘right to counsel’’ cases. American Political Science Review.
[5] Z. Hu, X. Li, C. Tu, Z. Liu and M.g Sun. 2018. Few-Shot Charge Prediction with Discriminative Legal Attributes. Proceedings of the 27th International Conference on Computational Linguistics, 2018.
[6] D. M. Katz, M. J Bommarito II, and J. Blackman. 2017. A general approach for predicting the behavior of the supreme court of the united states. PloS one, 2017.
[7] R Keown. 1980. Mathematical models for legal prediction. Computer/LJ 2:829.
[8] Q. Le and T. Mikolov. 2014. Distributed Representations of Sentences and Documents. PMLR 2014.
[9] W.-C. Lin, T.-T. Kuo, T.-J. Chang, C.-A. Yen, C.-J. Chen, and S.-d. Lin. 2012. Exploiting machine learning models for chinese legal documents labeling, case classification, and sentencing prediction. ROCLING, pages 140–141.
[10] C.-L. Liu and C.-D. Hsieh. 2006. Exploring phrase-based classification of judicial documents for criminal charges in chinese. ISMIS, pages 681–690.
[11] C.-L. Liu, C.-T. Chang, and J.-H. Ho. 2004. Case instance generation and refinement for casebased criminal summary judgments in chinese. JISE, 2004.
[12] B. Luo, Y. Feng, J. Xu, X. Zhang, and D. Zhao. 2017. Learning to predict charges for criminal cases with legal basis. In Proceedings of EMNLP, pages 2727–2736.
[13] T. Mikolov, I. Sutskever, K. Chen, G. S Corrado, and J. Dean. 2013. Distributed representations of words and phrases and their compositionality. NIPS 2013.
[14] Nagel SS. 1963. Applying correlation analysis to case prediction. Texas Law Review.
[15] O. M. Sulea, M. Zampieri, M. Vela, and J. V. Genabith. 2017. Exploring the use of text classi cation in the legal domain. In Proceedings of ASAIL workshop.
[16] G. Salton and C. Buckley. 1988. Termweighting approaches in automatic text retrieval. Information processing & management 24(5):513–523.
[17] C. Xiao, H. Zhong, Z. Guo, C. Tu, Z. Liu, M. Sun, Y. Feng, X. Han, Z. Hu, H. Wang, J. Xu.. CAIL2018: A Large-Scale Legal Dataset for Judgment Prediction.
[18] Sim Y, Routledge BR, Smith NA. 2015. The utility of text: the case of Amicus briefs and the Supreme Court. 29 AAAI conference on artificial intelligence.
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2020-07-01起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2020-07-01起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw