A Study on Legal Judgment Prediction using Deep Learning Techniques

Abstract

Legal Judgment Prediction (LJP) involves examining the given input case document and recommending the judgment prediction such as applicable law sections, charges, and penalties as delivered by the judge in the court. It assists the judges and lawyers in analyzing and resolving the given case. The various steps involved in LJP equip the lawyers with supporting points to argue the case in the court and the parties involved with the probability of winning the case by predicting the judgment outcome. This paper surveys recent state-of-the-art LJP algorithms published between 2018 and 2022 by focusing on various factors such as Deep Learning (DL) and Artificial Intelligence (AI) ambient techniques, civil and criminal case types, evaluation measures, various data sets available, prediction and modelling methods, challenges, and limitations. Based on this study we derived a taxonomy that will organize the collected papers into two channels called criminal and civil cases which are further classified based on the techniques used for prediction.