A Multi-domain Adaptation for sentiment classification algorithm based on Class Distribution.
Kongbing HuYuhong ZhangXuegang HuPublished in: GrC (2012)
Keyphrases
- classification algorithm
- domain adaptation
- class distribution
- sentiment classification
- class labels
- training set
- document classification
- concept drift
- labeled data
- unlabeled data
- training data
- test data
- sentiment analysis
- semi supervised
- naive bayes
- text classification
- semi supervised learning
- cross domain
- knn
- test set
- cost sensitive
- target domain
- k nearest neighbor
- support vector machine
- active learning
- data sets
- training examples
- supervised learning
- multiple sources
- training samples
- base classifiers
- decision trees
- classification accuracy
- transfer learning
- nearest neighbor
- co training
- feature selection
- machine learning
- multi label
- learning environment
- data analysis
- feature extraction
- text categorization
- prior knowledge
- knowledge discovery