| Online blogging continues to be a popular way for people to share and discuss their opinions with others in the community. Blogospheres also provide a forum where authors can follow other bloggers as well as recommend and leave comments on individual posts. Over time, certain authors become more popular with high numbers of recommendations and comments, leading to an increased influence on connected bloggers' sentiment and opinion within the network. Given the large number of users and issues being discussed in any blogosphere, it is extremely difficult to easily gauge patterns and trends in public sentiment and opinion, as well as any level of influence certain bloggers may have.;In this application of text mining, I introduce new techniques to model the evolving sentiment and opinion independently as well to track the blogger's influence within the blogosphere. While it may seem logical at first to consider sentiment as a form of opinion or derive an opinion based on the sentiment of the text, this analysis discovers numerous instances where authors either express subjectivity about a topic without any emotion, or where they incorporate sentiment words without any subjective viewpoints in their writings.;Additionally, this work also presents a social contagion model to represent a blogger's influence on neighboring authors within the blogosphere to determine where the overall level of sentiment and opinion converges to over time. This dissertation incorporates a modified epidemic model where sentiment and opinion represent infectious diseases and influence signifies an infection between individuals, differing from other recent work by incorporating the likely interactions of sentiment and opinion contagions within the same social network. |
