Using Case Facts to Predict Penalty with Deep Learning

With the promotion of Wisdom Court construction and the increasing completeness of judicial big data, the combination of judicial and artificial intelligence attracted more and more attention. The Judicial document is the most common textual information in cases. Due to the development of text analysis and processing techniques, we can mine more information from the judicial text and apply it in judgment. In this paper, we use the popular deep learning text classification algorithms to predict the imprisonment based on the fact of cases, which is expected to assist judges and staffs of procuratorate on sentencing. The result of our experiments shows the feasibility and utility of our method.