The Research On Key Technologies Of Network Public Opinion Monitoring Based On Cloud Computing

With the development of the society,Network public opinion,is the important form of public opinion,has been developing rapidly and influences the social reality.In our country,some network people or groups with ulterior motives anonymously comment on social sensitive topic,focus and hot topic by means of sudden,optional,hidden features of Network public opinion.And those guide the topic to the wrong direction and bring security hidden danger for the country's stability and unity.So it has become a hot issue to monitor network public opinion by using the computer technology of network public opinion and has strong practical significance.Based on this,This thesis studied on the hot spots founded in the key technology of Network public opinion monitoring system.The thesis mainly introduced some algorithm according to the research status of Network public opinion monitoring system and hot detection algorithm,such as the classic SinglePass algorithm,which can't efficiently process mass high-dimensional text or timely found potential Internet public opinion topic.This thesis also studied the features of Locality Sensitive Hashing and used SimHash algorithm to quickly find similar candidate data object from mass high-dimensional data object,so as to narrow the scope of the object and improve the efficiency of SinglePass algorithm clustering.In addition,according to the characteristics of the Storm platform and SimHash algorithm,this thesis write the Topology to distributed transform the single SimHash algorithm and designed an improved version of SinglePass hot topic detection algorithm.In order to verify the validity,this thesis designed verification experiment for SimHash algorithm.The results show that the algorithm can largely accurate to find the relevant documents,which proved that SimHash algorithm can obviously increase the clustering efficiency of SinglePass algorithm in the case of a similar clustering quality.In the same time,this thesis through experiments to verify the overall plan based on the Storm platform and the Topology of the distributed SinglePass algorithm.The results proved that the improved SinglePass hot topic detection algorithm can improve the efficiency of real-time data processing.All of experimental results can show that the Storm platform hot discovery mechanism based on SimHash algorithm from this thesis can ensure the accuracy of data processing and solve the problem of the efficiency of the traditional algorithm to improve the efficiency of network public opinion monitoring system.It can lay a foundation for the research and application of network public opinion monitoring system.