Estimating Feature Terms for Supporting Exploratory Browsing of Twitter Timelines

In recent years, use of the social networking/micro-blogging service Twitter has spread widely. Twitter users constantly post tweets and connect with an often huge number of other users. We need to optimize viewing a lot of tweets. For users who browse infrequently, many unread tweets appear on their timeline, making it difficult to view. It is also difficult to search the timeline for past tweets using the existing system. To solve these problems, we propose a system that supports exploratory browsing of Twitter timelines. The key element of the proposed system is that it presents feature terms from a specific time period, which enables users to see a relevant overview of their Twitter timelines. The experimental results show that our system can effectively extract feature terms from Twitter timelines within specific time periods.