VN-Legal-KG: Vietnam Legal Knowledge Graph for Legal Statute Identification on Land Law Matters

Legal Statute Identification (LSI) is a critical task within the realm of law, involving the identification of relevant statutory laws based on the natural language descriptions found in legal documents. Traditionally, this challenge has been approached as a single-class text classification problem. However, due to the inherent complexity of legal information, characterized by intricate connections and associations between various legal entities and concepts, we propose that a graph-based representation offers a more suitable and informative solution. In response to this need, our paper introduces VN-Legal-KG, an innovative Legal Statute Identification Knowledge Graph tailored to meet the specific requirements of Vietnamese users seeking clarity on Land Law matters. Leveraging cutting-edge graph neural network techniques, we also present a link prediction mechanism integrated into VN-Legal-KG, which addresses the LSI task as a multi-label classification problem, better aligning with real-world legal practices. Through experimentation with real-world data, our approach demonstrates favorable performance when compared to previous models reported in the literature.