基于微博的社会化媒体分析系统的设计与实现软件工程专业论文.docx

哈尔滨工业大学工程硕士学位论文 哈尔滨工业大学工程硕士学位论文 - - II - Abstract As the new media’s innovation and development, the microblog as the new forces of rapid development in the past two years, the number of internet users increased. Because microblogs popular on the internet and its greate user groups, many enterprises give it wide concern. Enterprises urgently need some products based on microblog to help them to do some related business development from the platform of microblog. Therefore, it is particularly important for the enterprise to monitor the microblog real-time. Based on the platform of sina microblog, this paper analyzes the advantages and disadvantages of the present existing related products on the market, and discusses the social medias analysis based on microblog applying the object-oriented program designing ideas. This paper argues that the existing monitoring analysis products for microblog mainly have two problems: the first is bad real-time data, it can not guarantee the systems data same as sina microblog platforms; The second is the precision of datas analysis, and the datas showed the user often have plenty of redundant information. According to the real-time problems of monitoring microblogs basic datas, the paper uses the distributed data acquisition engine based on Hadoop, and it gives effectively guarantees. At the same time, the distributed cluster environment speeds up the data processing speed in the great degree, and greatly improves the system’s performance. For the microblogs source information, it predicts bloggers behavior trends with the data analysis engine and the keywords analysis engine, and reduces the information redundancy. Through the rules engine and keywords highlighting technology etc, provides the classification mechanism for microblog datas containing specific keywords, and helps enterprise prompt understand users negative remarks and evaluation for certain products or their service. This paper based on the above technique, puts forwa