Research On Sensitive Information Identification Method Based On Sentiment Analysis

Since entering the information age,all kinds of good and bad information have flooded people's lives.Many overseas forces and criminals have distributed and disseminated some sensitive information through the Internet in order to incite and guide online public opinion.The various online media represented by Weibo contain a large number of texts on topics such as violence,terrorist attacks,political current events,etc.These texts reflect users' attitudes,views and tendencies towards the incident.Sensitive information identification has become an important research problem in recent years as an important means to avoid malicious guidance of Internet public opinion.This paper focuses on sentiment analysis and sensitive information identification methods in the task of identifying violent and sensitive information in web text.The main work of this article is as follows:(1)Aiming at the problem of sentiment analysis in sensitive text,a method of constructing sentiment analysis model for sensitive information recognition was proposed.In the model,in the traditional word2 vec semantic feature extraction method,the text semantic extraction method is improved,and the emotional features,relative position features of emotional words and sensitive words are extracted in the text.Combined with the two-way long-term and short-term memory model and selfattention mechanism,a sentiment analysis model(Sentiment Analysis Model For Sensitive Information Recognition,SAMFSIR)is obtained.This model analyzes the text and obtains three kinds of sentiment polarity.Experiments show that the method proposed in this paper is better than the existing methods in sentiment analysis.(2)Aiming at the task of identifying sensitive information in text,a sensitive information recognition method combining sentiment analysis is proposed.In this paper,by constructing sensitive triggering events and improving the text similarity algorithm based on the keyword part of speech,a text similarity algorithm combining part of speech(STEAP)is proposed.Then the SAMFSIR model and STEAP algorithm are used to calculate the sensitivity of the text.It is proved by experiments that the method has better recognition accuracy than traditional sensitive information recognition methods,and it also proves that text sensitivity and text sentiment are strongly related.(3)Aiming at the task of identifying sensitive information in web text,a sensitive information recognition system based on sentiment analysis was proposed.According to the method proposed in this paper,a sensitive information recognition system based on sentiment analysis is designed and implemented.According to the requirements,the system architecture and the functions of each module are designed and implemented,and the validity of the system is verified through use case testing.