Identifying Influential Citations in Scientific Papers.

The research described in academic papers builds on related existing work. This previous work can be considered as influential if it is used or extended in the current study or non-influential if it is only cited as background material. This study aims to identify influential citations, in particular examining the impact of sentiment associated with citations on citation influence. Our work leverages a set of statistical, contextual and metadata-based features to provide an assessment of the influence of cited works on the current work. Our findings show that sentiment analysis significantly enhances citation classification.