Design And Implementation Of Public Opinion Monitoring System Based On Cloud Platform

The rapid developments of information technology result in emergence of big data.In such an era with information rising fierce,the victor would be the people who can comprehensive and efficient grasp of the information.It is clear to see that accurately obtain valid information and timely realize the state of the information is crucial.The monitor system of public sentiment use rapid development technology of information to trim and analyze data,automatically,then it can build a valid,comprehensive and fast response monitor mechanism.It is very useful for government or enterprise to director event step by step to a well statement that through reading hot event and understanding the state of development of the event and then building correspondent policy.Based on filing information,reading a lot of literature,and according to the current development of the internet,we analyze,design and implement a monitor system for public opinion which is a cloud-based platform.Starts from the purpose and back ground of the monitor system of public opinion,here we analyze the demand of the system from economic,technology and operability via analyzing the function needed and features of the system,combining difficulty of the existing technology and implement of the system.Here we name it as Analyze.Design,on the basis of the demand of analysis,here we focused on the core of monitoring of public opinion,divided the function of the system reasonable,design the structure and modular system in a more detail and specific way,and design the system concerning to its security and management from the perspective of system.Implementation phase is the key point of the whole thesis.To implement the modules of the system separately which is based on the design,to achieve and analyze information more comprehensive and accurate,and make the system more flexible.Here we use the cloud platform,which is based on Hadoop,as the platform of information acquisition and analysis.As the calculation of MapReduce framework,which is based on Hadoop cloud platform,can put the task Map to each node to distributed execution,and finally all the sub tasks can be combined into the final result through Reduce.This mode is not only with high degree of parallelism but it can distribute tasks to machines at different locations evenly and enabling a more complete and accurate analysis data.This paper uses HDFS to store unstructured data,which can effectively store data and use the backup data of HDFS strategy,and ensure the security of the data.Meanwhile,through the deletion and extension of the nodes,it is easily to control processing ability of the system data.At the last,the testing results are given due to the test and verify of the system.