Maximum Entropy Models with Inequality Constraints A Case Study on Text Categorization.
Data sparseness or overfitting is a serious problem in natural language processing employing machine learning methods. This is still true even for the maximum entropy (ME) method, whose flexible modeling capability has alleviated data sparseness more successfully than the other probabilistic models...
| Gepubliceerd in: | Machine learning. 60, 1-3 (2005). |
|---|---|
| Hoofdauteur: | |
| Formaat: | Artikel |
| Taal: | English |
| Onderwerpen: |