Levenberg-Marquardt Learning Neural Network For Part-of-Speech Tagging of Arabic Sentences

Author:

Abstract: Part-of-speech tagging is usually the first step in linguistic analysis. Also, it is a very important intermediate step to build many natural language processing applications. This paper examines the application of neural networks to the task of tagging Arabic sentences. The network is trained with the help of Levenberg-Marquardt learning algorithm. Corpora of 24,810 words are collected and manually tagged to train the neural networks and to test the performance of the developed POS-Tagger. The developed tagger achieved an accuracy of 98.83% when evaluated on the train set and 90.21% on the test set. The performance of the Levenberg-Marquardt learning algorithm was compared with the performance of the traditional Backpropagation learning algorithm. It was found that the Levenberg-Marquardt Learning neural network is an efficient approach and more effective than the traditional Backpropagation learning algorithm for tagging Arabic words.

WSEAS Transactions on Computers, ISSN / E-ISSN: 1109-2750 / 2224-2872, Volume 13, 2014, Art. #25

PDF
XML
Cite
×
Citation Tools
WSEAS AMA
Hasan Muaidi, "Levenberg-Marquardt Learning Neural Network For Part-of-Speech Tagging of Arabic Sentences," WSEAS Transactions on Computers, vol. 13, pp. 300-309, 2014, DOI:
Hasan Muaidi. Levenberg-Marquardt Learning Neural Network For Part-of-Speech Tagging of Arabic Sentences. WSEAS Transactions on Computers. 2014;13:300-309.
Copy to Clipboard
Citation copied to Clipboard

Certification

---