| With the development of mobile Internet technology,smart terminals and social software,network big data has become the most representative big data.Economic,cultural,military and social activities are strongly dependent on the Internet.Proper application of network big data can play a positive role in maintaining social security and stability.Many public security incidents will keep track of the Internet in the process of development.Therefore,it is of great significance to study the hot event extraction and public safety event grading in the network big data environment to maintain social stability and prevent the spread of public opinion.Firstly,by analyzing the text data in the social network,calculate the heat score,merge the synonyms,calculate the growth trend,and extract the popular words from the network big data text.Further use the phrase mutual information and word frequency screening events to express more powerful phrases,identify and extract hot events from the network big data.Then,in combination with the characteristics of public safety events in the network,three indicators for evaluating public safety events are defined: heat,emotion and sensitivity,so that public safety events are marked out from the hot events that have been extracted.On this basis,public safety-related data and non-public security data are stored separately using a hybrid storage system containing SSDs and mechanical hard disks to solve the problem of efficient data query.Experiments show that by comparing the public safety event rating method with other public opinion discovery algorithms,the rating method can improve the accuracy rate of public safety event recognition by 100.5%,the recall rate by 6.9%,and the comprehensive evaluation coefficient of accuracy and recall rate F1.The value is increased by 45%.In the subsequent query access,the hybrid storage system is more than 16 times higher than the full use of SSD storage,which proves the effectiveness of the research scheme. |
PDF Full Text Request