Research on intelligent legal decision support system based on big data analysis

Abstract

Abstract This paper constructs an intelligent decision support system using text combination feature extraction, case similarity, XBoost sentencing prediction model, and other related technologies. The performance of the smart decision support system is tested for grabbing responses and monitoring response reliability. The XGBoost algorithm is used to construct the sentencing prediction model to achieve the intelligence of sentencing prediction. The effectiveness of the sentencing prediction model is examined by comparing the XGBoost model with different algorithms (Random Forest, CBDT, CNN). The practical application of intelligent decision support systems was summarized, highlighting the positive and negative effects. The results show that the overall responsiveness performance of the intelligent decision support system’s magisterial document capture and retrieval is reliable, and the number of captured and retrieved magisterial documents rises as the capture time increases. In XGBoost’s sentencing prediction, nine overlapping parts of the ten keywords were extracted for sentence and fine. In contrast, the keyword similarities are all higher than 0.5, and the difference between the predicted and real sentence values and fine is small. The intelligent decision support system has resulted in a gradual decrease in the number of court cases and an improvement in efficiency. The re-sentencing and retrial initiation rate decreased by over 20% from 2019 to 2023.