Arabic Broken Plural Model Based on the Broken Pattern

Broken (or irregular) plurals present a significant issue in natural language processing. It is known that there is not a fixed and general rule assembling all the broken plural forms. Extracting the plurals or reducing them to their singular form is a task that affects the performance of information retrieval, annotation, tagging tasks, and many other NLP applications. This paper describes, on one hand, our own Arabic broken plural list and, on the other hand, the process of extracting the plural and its singular form using several machine learning classifiers. Obtained results show that the Random Forest classifier outperforms the other statistical classifiers with an accuracy of approximately 98%.