Verdict Prediction for Indian Courts Using Bag of Words and Convolutional Neural Network

All courts in India publish judgment by statistically analyzing the data of different cases and understanding the verdict from precedents judgments and the statute law. The approach of our study is to highlight the importance of Convolutional Neural Network (CNN) and Natural Language Processing (NLP) in the legal domain. The Bag of Words technique is being proposed i.e., one of the NLP tool to analyze the text of the court proceedings to extract the keywords from the text and CNN to classify each case into its charges (as per judicial law of India), to predict whether it is a bailable or a non-bailable offence and to give an approximate judicial decision. The results show that this method has an average accuracy of 85% in prediction based on the IPC (Indian Penal Code) which is extracted from the case files. This research work data is taken from the judicial pronouncement and the constitution of India.