| 1 |
An Algerian dialect: Study and Resources
|
|
|
|
In: ISSN: 2158-107X ; EISSN: 2156-5570 ; International journal of advanced computer science and applications (IJACSA) ; https://hal.archives-ouvertes.fr/hal-01297415 ; International journal of advanced computer science and applications (IJACSA), The Science and Information Organization, 2016, 7 (3), pp.384-396. ⟨10.14569/IJACSA.2016.070353⟩ (2016)
|
|
BASE
|
|
Show details
|
|
|
| 2 |
Grapheme To Phoneme Conversion - An Arabic Dialect Case
|
|
|
|
In: Spoken Language Technologies for Under-resourced Languages ; https://hal.inria.fr/hal-01067022 ; Spoken Language Technologies for Under-resourced Languages, May 2014, Saint Petesbourg, Russia (2014)
|
|
BASE
|
|
Show details
|
|
|
| 3 |
Building Resources for Algerian Arabic Dialects
|
|
|
|
In: 15th Annual Conference of the International Communication Association Interspeech ; https://hal.inria.fr/hal-01066989 ; 15th Annual Conference of the International Communication Association Interspeech, ISCA, Sep 2014, Singapour, Singapore (2014)
|
|
BASE
|
|
Show details
|
|
|
| 4 |
Evaluation of Topic Identification Methods on Arabic Corpora
|
|
|
|
In: ISSN: 0972-7272 ; Journal of Digital Information Management ; https://hal.inria.fr/hal-01586544 ; Journal of Digital Information Management, Digital Information Research Foundation, 2011, 9 (5), pp.8 double column (2011)
|
|
Abstract:
International audience ; Topic Identification is one of the important keysfor the success of many applications. Indeed, there are fewworks in this field concerning Arabic language because oflack of standard corpora. In this study, we will provide directlycomparable results of six text categorization methods on anew Arabic corpus Alwatan-2004. Hence, Topic UnigramLanguage Model (TULM), Term Frequency/Inverse DocumentFrequency (TFIDF), Neural Network, SVM, M-SVM and TRhave been experimented, and showed that TR-Classifier isthe most efficient among the set of classifiers, nevertheless,only binary SVM outperformed it thanks to its characteristics.Moreover, we should note that the size of Alwatan-2004 corpusused to achieve our experiments is considered the mostimportant compared to any other Arabic corpus which hadbeen used for topic identification experiments until now. Inaddition, we aim through using small sizes of vocabularies toreduce the time of computation. This is important for adaptivelanguage modeling, particularly Topic Adaptation, which isrequired in real time applications such as speech recognitionand machine translation systems. Our experiments indicatethat the results are better than other works dealing with Arabictext categorization.
|
|
Keyword:
[INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL]; Arabic Language; SVM; Topic Identification
|
|
URL: https://hal.inria.fr/hal-01586544
|
|
BASE
|
|
Hide details
|
|
|
| 5 |
TR-Classifier and kNN Evaluation for Topic Identification tasks
|
|
|
|
In: ISSN: 0973-5836 ; International Journal on Information and Communication Technologies ; https://hal.inria.fr/hal-01586549 ; International Journal on Information and Communication Technologies, Serials Publications, 2010, 3 (3), pp.10 (2010)
|
|
BASE
|
|
Show details
|
|
|
|
|
