What makes a tweet relevant for a topic
Ke Tao, Fabian Abel, Claudia Hauff, Geert‐Jan Houben · 2012
Users who rely on microblogging search (MS) engines to find relevant microposts for their queries usually follow their interests and rationale when deciding whether a retrieved post is of interest to them or not. While today’s MS engines commonly rely on keyword-based retrieval strategies, we investigate if there exist additional micropost characteristics that are more predictive of a post’s relevance and interestingness than its keyword-based similarity with the query. In this paper, we experiment with a corpus of Twitter messages and investigate sixteen features along two dimensions: topicdependent and topic-independent features. Our in-depth analysis compares the importance of the different types of features and reveals that semantic features and therefore an understanding of the semantic meaning of the tweets plays a major role in determining the relevance of a tweet with respect to a query. We evaluate our findings in a relevance classification experiment and show that by combining different features, we can achieve a precision and recall of more than 35 % and 45 % respectively. 1.