Qillah: A Morphological Extension for Identifying Plural-of-Paucity Arabic Words

Abstract

Plural of paucity is one type of broken plural used in the classical Arabic. It is used when the number of people or objects ranges from three to 10. Based on our evaluation of four current state-of-the-art Arabic morphological analyzers, there is a lack of identification of broken plural words, specifically the plural of paucity. Therefore, this paper presents Qillah (paucity), a morphological extension that is built on top of other morphological analyzers and uses a hybrid rule-based and lexicon-based approach to enhance the identification of plural of paucity. Two versions of the Qillah were developed, one is based on FARASA morphological analyzer and the other is based on CALIMA Star analyzer, as these are some of the best-performing morphological analyzers. We designed two experiments to evaluate the effectiveness of our proposed solution based on a collection of 402 different Arabic words. The version based on CALIMA Star achieved a maximum accuracy of 93% in identifying the plural-of-paucity words compared to the baselines. It also achieved a maximum accuracy of 98% compared to the baselines in identifying the plurality of the words.