Court Similar Case Recommendation Model Based on Word Embedding and Word Frequency

Mining valuable services from the massive big data of court judgment texts is the focus of current judicial informatization research. The paper proposed a novel TF-W2V comprehensive similarity model based on deep neural network to calculate the similarity of judgment texts.The proposed approach can implement semantic retrieval of judgment texts similar with cases that the users want to search, and recommend the critical information of similar cases needed by courts in the process of handling cases, which is helpful for the new case decision-making. Experimental results show that the accuracy of the proposed model is up to 37% higher than other models.