This paper reports on our ongoing work on the analysis of sentiments, i.e., individual and collective stances, in Hansard (Hansard is a publicly available transcript of UK Parliamentary debates). Significant work has been carried out in the area of sentiment analysis particularly on reviews and social media but less so on political transcripts and debates. Parliamentary transcripts and debates are significantly different from blogs and reviews, e.g., the presence of sarcasm, interjections, irony and digression from the topic are commonplace increasing the complexity and difficulty in applying standard sentiment analysis techniques. In this paper we present our sentiment analysis methodology for parliamentary debate using known lexical and syntactic rules, word associations for the creation of a heuristic classifier capable of identifying sentiment carrying sentences and MP stance.
展开▼
