Target-dependent sentiment classification in news articles reporting on political events. Includes a high-quality data set of over 11k sentences and a state-of-the-art classification model. - View it on GitHub