Storage Management Method For Public Safety Events Based On Clustering And Hot-Cold Data Classification

Public safety events are related to the national economy and the people's livelihood,and timely perception and processing of public safety events is becoming increasingly important in the information age.Massive social network data contains public security event information and non-public security event information,which are mixed together,making querying difficult.How to efficiently store and query a public safety event information is a huge challenge.At the same time,the event is constantly evolving and evolved.How to fully exploit the access locality of event information to store and manage event information is of great significance for accurate and real-time query analysis of events.A public security event storage management method based on clustering and hot-cold data classification is designed and implemented in this paper.On the one hand,the community partitioning algorithm is optimized using local modularity to correlate the acquired network information.According to the different types of datasets,the semantic distance between different text nodes is calculated for the news webpage text data,and the event clustering is completed by the optimized community partitioning;for the microblog text data,the keyword provenance graph is constructed by using the information entropy.Then the events are clustered by community division for the graph,which improves the event clustering accuracy and reduces the time overhead.On the other hand,using the principle of access locality of event information,the clustered event information is stored centrally,and the performance advantage of the solid state disk compared with hard disk driver is used to separate and store the event information according to the hot and cold,so as to further improve the query performance of public safety event?The experimental results show that the event cluster algorithm proposed in this paper can get cluster purity over 94% in numerical data while applied to public safety events expressed by long texts of news webpage and short-pages of sina weibo.And the accurate query time reduced over 52% in comparison with the original data.The system proposed is proved to be efficient.