There is high demand for computational tools that can automatically label tweets (Twitter messages) as having positive or negative sentiment, but great effort and expense would be required to build a large enough hand-labeled training corpus on which to apply standard machine learning techniques. Going beyond current keyword-based heuristic techniques, this paper uses emoticons (e.g. ':)' and ':(') to collect a large training set with noisy labels using little human intervention and trains a Maximum Entropy classifier on that training set. Results on two hand-labeled test corpora are compared to various baselines and a keyword-based heuristic approach, with the machine learned classifier significantly outperforming both. / text
| Identifer | oai:union.ndltd.org:UTEXAS/oai:repositories.lib.utexas.edu:2152/ETD-UT-2011-08-3823 |
| Date | 02 February 2012 |
| Creators | Speriosu, Michael Adrian |
| Source Sets | University of Texas |
| Language | English |
| Detected Language | English |
| Type | thesis |
| Format | application/pdf |
Page generated in 0.0025 seconds