The Research Of Enterprise Online Reputation Based On The Text Sentiment Analysis

With the rapid development of Internet, the online communities become more and more popular, such as blog, BBS and microblog. Internet has become a place for everyone to comment and share personal feelings. While for the enterprises, we can make good use of these data. For example, we can dig out the status and trend of enterprise online reputation.This paper makes comprehensive use of case analysis, content analysis, statistical analysis and other research methods to make research on how to collect data about netizens’ comments and feelings from blog, BBS and microblog community, with the use of modern mature text sentiment analysis technology, to analyze and mining the netizens’ emotions polarity hiding in their comments and attitudes towards the products and services of the enterprise. So that in the end we can build a set of online reputation characteristics analysis method, the enterprise online reputation evaluation system and early warning system, to help the enterprise achieving an online reputation analysis and early warning in current situation.In this paper, the main research content and innovation points are as follows:(1) To build a set of enterprise online reputation metadata collector, including a focused crawler for blogs and BBS data and a microblog data collector for SINA Weibo. This collector can obtain, format extract and store enterprise products and the service metadata in the blogs, BBS and SINA Weibo.(2) To build a corporate reputation online analysis toolkit. The toolkit use LingPipe to analyze netizens’ attitudes, and the features of enterprise online reputation from the aspect of overall trend, age structure of the netizens, regional distribution, education background, the influence on the network and authentication type, etc. Meanwhile this paper presents an evaluation indicator system and warning indicator architecture of enterprise online reputation, which can realizes the current situation analysis and warning of the online reputation. This is the innovation points in this paper.(3)This paper demonstrate the effectiveness of the metadata collector and analysis toolkit by two instances of Mengniu Dair and Siemens refrigerators, thus proving the research results has a certain reference value and guiding significance to the enterprise.The innovation points of this paper are to analyze the features of enterprise online reputation from six aspects. Based on these features, this paper presents an evaluation indicator system and warning indicator architecture of enterprise online reputation.