word-based and dependency-based method

Online political discussions have received a lot of attention over the past years. In this paper we compare two sentiment lexi-con approaches to classify the sentiment of sentences from political discussions. The first approach is based on applying the number of words between the target and the sentiment words to weight the sen-tence sentiment score. The second ap-proach is based on using the shortest paths between target and sentiment words in a dependency graph and linguistically mo-tivated syntactic patterns expressed as de-pendency paths. The methods are tested on a corpus of sentences from online Nor-wegian political discussions. The results show that the method based on depen-dency graphs performs significantly better than the word-based approach.