摘要
Abstract
Sensitive information in digital media networks has the characteristics of diversity and concealment,and there are many vocabulary residues that affect the accuracy of annotation and recognition.A sensitive information automatic annotation method in digital media networks is proposed based on deep neural networks and natural language processing technology.A sensitive information dataset is constructed by integrating natural language processing technology,which includes text features related to sensitive information in digital media networks.A information recognition model based on deep neural networks is es-tablished,and sensitive information is classified and automatically annotated.The experimental results show that in practical applications,the retention stability and integrity of the dataset are high,and sensitive information with relatively complex infor-mation composition has strong classification accuracy and precision.Overall,there is no occurrence of oblique classification,and the overall performance is better,with significant advantages.关键词
深度神经网络/敏感信息/自动标注/自然语言处理技术/数字媒体Key words
deep neural networks/sensitive information/automatic annotation/natural language processing technology/digital media分类
电子信息工程
