@article{347,
  author = {Al-Gaphari.G., Al-Nuzaili A},
  title = {Arabic Text Mining Using Maximum Entropy Model},
  journal = {International Journal of Computational Linguistics Research},
  year = {2010},
  volume = {1},
  number = {2},
  doi = {},
  url = {http://www.dline.info/jcl/fulltext/v1n2/2.pdf},
  abstract = {Building a high performance classifier requires an efficient training algorithm as well as a high performance testing algorithm .With the present effort, we propose to focus on the development of an automated Maintainable information classifying system as a main goal. The system has two phases: learning phase and testing phase. On the one hand, the system accepts a set of Arabic classified documents as a real training data set, during its learning phase. The system learning technique based on the so- called Maximum Entropy Model. The model enables the system to learn the parameters weights. On the other hand, the system accepts, during its testing phase, any randomly selected Arabic unclassified document or documents. Then it uses the estimated weights, learned so far, to decide whether that document or documents belongs or belong to one or more predefined categories, depend on their context. The maximum entropy model was implemented. Hence, making a conclusion that emphasizes the model efficiency for Arabic text categorization in terms of learning speed, real time classification speed, and classification accuracy.},
}