Scary films good, scary flights bad: topic driven feature selection for classification of sentiment.
Scott NowsonPublished in: TSA@CIKM (2009)
Keyphrases
- feature selection
- classification accuracy
- support vector
- text classification
- classification models
- feature set
- machine learning
- decision trees
- feature space
- support vector machine
- classification algorithm
- model selection
- image classification
- feature selection algorithms
- sentiment analysis
- unsupervised learning
- opinion retrieval
- feature selection and classification
- feature extraction
- classification performances
- feature reduction
- irrelevant features
- method for feature selection
- polarity classification
- accurate classification
- small sample
- sentiment classification
- knn
- multi class
- mutual information
- text categorization
- feature subset
- classification method
- high dimensionality
- microarray datasets
- social networks
- topic models
- classification rate
- bayes classifier
- feature ranking
- discriminative features
- ensemble classifier
- selected features
- opinion mining