Simpler is better? Lexicon-based ensemble sentiment classification beats supervised methods.
Lukasz AugustyniakTomasz KajdanowiczPiotr SzymanskiWlodzimierz TuliglowiczPrzemyslaw KazienkoReda AlhajjBoleslaw K. SzymanskiPublished in: ASONAM (2014)
Keyphrases
- sentiment classification
- supervised methods
- text classification
- labeled data
- sentiment analysis
- labeled training data
- cross domain
- semantic segmentation
- semi supervised
- weakly supervised
- methods require
- opinion mining
- word sense disambiguation
- feature selection
- transfer learning
- training data
- unlabeled data
- labor intensive
- semi supervised learning
- object class
- domain specific
- cross lingual
- text mining
- training set
- text categorization
- topic models
- data mining
- text classifiers
- data sets
- probabilistic model
- bayesian networks
- information retrieval