DEVELOPING CASE-BASED REASONING SYSTEM FOR CIVIL LAW DECISION-MAKING

Abstract

Law is a system created by a government to control and regulate people's behavior in a country. Social control refers to the rules or means that allow individuals and groups to behave in certain ways. The government controls and monitors what individuals and groups do with the help of formal public authorities such as courts and police. Addressing gaps in civil law decision making, such as ethical and bias considerations, capacity interpretation and transparency, privacy, and data security, integration with legal practices current, impacting employment and legal training, as well as the legal and regulatory framework for AI in law. This research aims to create case-based reasoning systems that can facilitate the decision-making process in civil cases. To identify problems in a specific field and develop a model, the researcher used the design science research design method. The researcher used a variety of methods to gather knowledge, such as conducting interviews, researching documents, and analyzing data using Waikato. In this study, the researcher used a predictive data mining approach to organize data and identify typical examples, with an emphasis on a technique called classification. To analyze and test the data collected from the court, the researcher used a data mining tool called Waikato Environment for knowledge analysis. Three experiments were performed using three different classification algorithms, namely j48, partial, and naive, to determine the best model and select the best performing classification algorithm in data mining. After conducting experiments, the researcher concluded that the J48 classification algorithm performed better than other classifiers in generating cases for a case-based system. Using jCOLIBRI version 1. 1, the researcher built a model that was tested with different samples. The developed prototype case-based reasoning system is then evaluated for its performance using system testing and user acceptance testing. System testing showed 93% recall, 84% precision, and 87% F-measure. Professional users and domain experts and 89 have also tested this system. 2% of them have accepted it. This research has contributed to improving artificial intelligence systems in the legal field and provided valuable information for future research and practical use