点击次数：

论文类型：会议论文

发表时间：2018-01-01

收录刊物：CPCI-S

页面范围：132-137

关键字：deep learning model; text clustering; the deep CFS model

摘要：With the fast development of the Internet technology, the court text information is collected from various fields at an unprecedented speed, such as Weibo and Wechat. This big court text information of high volume poses a vast challenge for the judge making reasonable decisions based on the vast cases. To cluster the reasonable assistant cases from the vast cases, we propose a deep CFS model for the text clustering, which can cluster the court text effectively, in this paper. In the proposed model, a robust deep text feature extractor is designed to improve the cluster accuracy, in which an ensemble of deep learning models are used to learn the deep features of the text. Furthermore, the CFS algorithm is conducted on the extracted deep text features, to discover the non-spherical clusters with the automatic find of the cluster centers. Finally, the proposed deep cluster model is evaluated on two typical datasets and the results show it can perform better than compared models in terms of the cluster accuracy.