A Study of the Effectiveness of Using Deep Learning Algorithms to Analyze Legal Risk Identification in Social Work Programs

In recent years, the state has vigorously promoted the purchase of public services from social forces by the government to support the participation of social organizations in social work service projects. In the context of social work programs get attention, the law as an important part of the judicial protection system has an important impact on the implementation of social work programs, therefore, the research on legal risk identification for social work programs should not be ignored. This paper elaborates the principle of deep learning, lays the foundation for legal risk identification based on deep learning, and constructs a deep neural network risk identification model based on DNN model using pytorch framework. In model training, after stabilization of the model, the loss value is reduced to below 0.2, the accuracy rate reaches approximately 98%, and the model performance is good. On this basis, in order to further study the effectiveness of the model in legal risk identification, through the comparison with other models, the legal risk identification model in this paper is the best in the three indexes of precision rate, recall rate and F1 value, which are 89.32%, 86.23% and 87.35% respectively, and the average accuracy of the six legal risk identification is 88.152%. The legal risk identification model based on deep learning in this paper has obvious improvement in identification effect compared with other models, verifies its effectiveness in legal identification of social work programs, and has good guidance for specific applications in the legal field of social work programs.