Treffer: Scaling Up: Analyzing Classification Outcomes in Large Korean Legal Document Datasets

https://doi.org/10.21860/j.15.2.12 This paper proposes an integrated approach that combines artificial intelligence models for automatic classification and prediction of Korean legal judgments. Given the complexity of the Korean legal system and the diversity of its legal issues, this study utilizes a transformer-based model, specifically, KLUE/BERT and KLUE/RoBERTa, to classify and predict legal judgment documents. By leveraging these models, this study addresses the challenges posed by the intricate legal language and diverse topics within Korean legal documents, significantly improving the efficiency and accuracy of classification tasks. The proposed approach enhances the automation and reliability of legal document predictions, demonstrating exceptional performance in managing the complexities of legal language. Specifically, the models facilitate a deeper understanding of the context of Korean legal judgments, thereby increasing the reliability of the prediction results. Moreover, this study introduces a novel integrated framework that significantly enhances the performance of automated legal document processing and prediction systems. This framework supports legal consultations, document management, and automated judgment systems, representing a significant advancement in the application of artificial intelligence in the legal domain.