PS-LSTM：中文隐私敏感实体自适应识别模型

his paper proposes a Chinese privacy sensitive adaptive recognition model CPS-LSTM, which is based on the character representationconstruction method of fused word vector. The model has a more streamlined network structure, has great advantages in statement processing performance, and can support parallel processing of multiple statements. The experiment shows that compared with the existing NER models, this model has faster convergence speed under the same data set and super-parameter combination, and can mine more potential private data entities from the target text, with better entity recognition accuracy.