Enhancing Legal Decision Making: WRIT Case Outcome Prediction with LegalBERT Embeddings and AdaBoost Classifier

Legal decision-making, the process of determining the outcome of court cases, is a pivotal aspect of the justice system. It involves evaluating evidence, interpreting laws, and ensuring justice is served. However, this process is challenging due to the complexity of legal cases, varying interpretations of laws, and the potential for lengthy court proceedings. In addition to these reasons, India also has one of the world's lowest ratios of judges to population. For swift and efficient disposal of justice, it is important to fast-track this process as much as possible. A traditional solution to this problem would take a considerable amount of time to show any meaningful results considering the current backlog of cases and the rate at which new cases are being filed each day. In order to overcome these difficulties, we employ text embeddings from LLM's with statistical classifiers to effectively and precisely forecast the outcomes for writ petitions. Our approach delivers an accuracy of 54% and a F1-score of 0.42 with LegalBERT embedder and AdaBoost classifier.