Automatic Extraction of Legal Citations using Natural Language Processing

Abstract

The accessibility of legal documents to the different actors of the judicial system needs to be ensured for the implementation of a strong international rule of law. The gap of such accessibility is being addressed by the Jus Mundi multilingual search-engine for International Law. The data updated on this platform is qualified by skilled lawyers. However, the interconnection of references within such documents, is a key feature for lawyers since, a major part of the legal research is analysing such citations to support their arguments. The process of interconnecting such references can prove to be expensive as well as time-consuming, if completed manually. Hence, the purpose of this research is to automatically extract such legal citations within international law, using Natural Language Processing (NLP), enabling the interconnectivity of documents on Jus Mundi. This study also discusses and addresses research gaps within this subject, especially in the domain specific to International Law. The method followed to achieve the automation is building an adaptable model through Regular-Expression based annotation language named JAPE (Java Annotation Patterns Engine). This set of automatically extracted links are then to be integrated with the search engine, having direct implication in the enablement of smoother navigation, making the law more accessible. This research also contributes to the state of the art bringing closer the eventual use of NLP in applications used to interact with International Law documents.