基于深度学习的涉密敏感信息识别技术研究OA

To improve the intelligence level of sensitive information management work,this paper proposes a BERT-BGRU-CRF Deep Learning method to achieve automatic recognition of sensitive information.This method first preprocesses the text information using the BERT model,then uses the Bidirectional Gated Recurrent Unit(BGRU)model to obtain contextual semantic features,and finally inputs the extracted information into the Conditional Random Field model for sequence annotation to obtain the optimal solution.The experimental results show that on the self-built dataset,the proposed method achieves higher scores in accuracy,recall,and F1 value compared to the three recognition methods BERT-CRF,BERT-LSTM-CRF,and BERT-BiLSTM-CRF,proving that this method is suitable for intelligent identification of sensitive information.