Infer information from Tweets. Useful for human-centered computing tasks, such as sentiment analysis, location prediction, authorship profiling and more!
For positive-negative-neutral classification, compare having a single classifier that can output any label against a two level hierarchical classifier that will first determine if it is neutral (objective), and if not then pass to a second classifier that determines positive or negative.
For positive-negative-neutral classification, compare having a single classifier that can output any label against a two level hierarchical classifier that will first determine if it is neutral (objective), and if not then pass to a second classifier that determines positive or negative.