A transductive extension of maximum entropy/iterative scaling for decision aggregation in distributed classification.
David J. MillerYanxin ZhangGeorge KesidisPublished in: ICASSP (2008)
Keyphrases
- maximum entropy
- iterative scaling
- class conditional
- maximum entropy principle
- text classification
- support vector
- minimum cross entropy
- supervised learning
- support vector machine
- principle of maximum entropy
- conditional random fields
- transformation based learning
- classification accuracy
- maximum entropy discrimination
- decision trees
- support vector machine svm
- decision making
- semi supervised learning
- unsupervised learning
- image classification
- svm classifier
- labeled data
- kernel logistic regression
- model selection
- feature vectors
- training set
- information retrieval