Design and Implement an Automatic Neural Tagger Based Arabic Language for NLP Applications
Jabar, H. Yousif,
To this day, various methods such as the statistical model, rule-based models and Support Vector Machine have been used to implement the POS tagger systems. However, these approaches require a large amount of data in order to adapt and implement the POS tagger. The neural approaches, on the other hand, only use lesser amount of data to perform the training and learning stages. The Arabic part of speech (POS) based multilayer perceptron is designed and implemented, while the Error back-propagation learning algorithm is used. The experiments have proven that not only the multilayer perceptron tagger is highly accurate (of 99.99%), it requires low processing time and a lesser amount of data to achieve the learning phase. A new coding scheme is investigated and implemented.