Necessity of Feature Selection when Augmenting Tweet Sentiment Feature Spaces with Emoticons.
Joseph D. PrusaTaghi M. KhoshgoftaarAmri NapolitanoPublished in: FLAIRS Conference (2016)
Keyphrases
- sentiment analysis
- feature space
- feature selection
- sentiment words
- text classification
- dimensionality reduction
- opinion mining
- high dimensionality
- feature set
- text categorization
- sentiment classification
- high dimensional
- positive and negative
- sentiment lexicon
- classification accuracy
- sentence level
- image representation
- training set
- mutual information
- information gain
- support vector machine
- feature vectors
- text mining
- low dimensional
- micro blog
- kernel function
- bag of words
- feature selection algorithms
- machine learning
- topic specific
- feature subspace
- positive or negative
- data sets
- model selection
- natural language processing
- feature extraction
- social media
- feature subset
- social networks
- text messages
- product features
- irrelevant features
- support vector
- product reviews
- data points
- multi class
- unsupervised learning
- nearest neighbor search
- selected features
- image retrieval
- kernel matrix
- multiscale
- multiple features
- user generated content
- hyperplane
- polarity classification