A Social Media Harvester Based On RSSBus

With the development of the Internet, more and more people began using theInternet in their lives, Internet has changed our lives to bring conveniences rather thanset a barrier. Currently, Social Networking Services (SNS) is the most popular one of allthe applications on the Internet. SNS provides us a better tool and platform for sharinginformation or making a discussion with other people. The amount of information inSNS is increasing with high quality year by year, so it is necessary to establish acollection and analysis system to help researchers to collect the messages easily. Thatwill excavate out the potential and the valuable information from SNS.In this paper, we are using RSSBus to build a social media harvester system (SMH).RSSBus supports user to create web application on its platform, and user can invoke theservices by multiple protocols. SMH will help users to collect messages from Twitter bykeywords. Compare to other SNS applications, information from Twitter will not bedelayed and losses. SMH supports user to customize the collection processes, likechoosing the keywords, searching cites and so on. The collection tasks are controlled bya scheduler function, it will start the processes in time, and adjusts the time interval ofeach process.From the data collected we can use statistical method to find out the relationbetween the keywords and the collected data. In this paper, we choose a multipleregression model in MATLAB to predict the box office income by the amount of thecollected tweets, and the result is acceptable. The result will provide a reference to thefilm company. Currently, this system has been running for more than two months, andhas collected19movies information. In addition, there are23movies in the collectionjob. In general SMH is a distributed system, the models of this system have beendeployed in different servers and invoke each other by web service.