Methods

To develop the Arabic neural network tagger, students and the faculty member must do
the following:

• Investigate current research activity in the field of natural language processing
  and more specifically, POS tagging.
  
• Read about neural networks. We will need to decide the type of network that
  we will use as well as determine the best training procedure.
  
• The tagger will be developed under the Windows operating system since this
  supports the Arabic language and the Unicode encoding format.
  
• The tagger will be developed in C++ using Microsoft Visual C++.
  
• The neural network tagger will be trained on a manually tagged corpus that
  was compiled by the faculty member while developing the statistical POS
  tagger.
  
• A comparison of the results obtained by the neural network tagger and the
  statistical tagger will be performed.