Abstract
Twitter becomes a popular microblogging platform that let people to express their opinions on the web in recent years. Companies with new products always want to find consumers or public opinions about their products and services after they released new products. Due to this reason, more and more researchers use the opinion mining technology on this microblog media that contains abundant and real-time information, to extract useful opinions and information. In this paper, we aim to mine the opinions on Twitter and further extract the competition relations discussed on Twitter. For Example, if we want to know how people express their opinion about “Packer” (an American football team name), we also want to know what the Packer’s competitors are. In this paper, we introduce a hashtag graph and use the ranks in this graph to represent the competition behavior and competition components (competitors).