The Use of Text Similarity and Sentiment Analysis to Examine Rationales in the Large-Scale Online Deliberations
Wanting Mao, Lu Xiao, Robert E. Mercer · 2014
To overcome the increasingly time consuming and potentially challenging identification of key points and the associated rationales in large-scale online deliberations, we propose a computational linguistics method that has the potential of facilitating this process of reading and evaluating the text. Our approach is novel in how we determine the sentiment of a rationale at the sentence level and in that it includes a text similarity measure and sentence-level sentiment analysis to achieve this goal.