Mizan: An Innovative Tool for Building Arabic Morphological Pattern Dictionaries and Enhancing Lexical Resource Development

Abstract

Mizan is a computational tool designed to enhance the development of Arabic morphological pattern dictionaries and facilitate the creation of advanced lexical resources for linguistic research. By systematically identifying and classifying morphological patterns (awzān), this tool automates the extraction of words sharing the same patterns, and provides an effective solution to overcome the complex challenges of Arabic morphology. This innovation extends beyond the standard pattern-based analysis to be able to process extensive Arabic corpora in a time-efficient manner, and handle datasets larger than five hundred million words with high precision. Addressing a critical gap in Arabic computational linguistics, the system introduces a novel approach to extract and analyze different morphological patterns. Unlike traditional manual methods or partial solutions available for linguistic research, Mizan offers detailed morphological statistics and promotes the development of comprehensive pattern-based dictionaries to achieve high accuracy. Using precision and recall metrics, empirical evaluations confirm its reliability and effectiveness in recognizing complex linguistic phenomena, including non-standard word forms and rare morphological patterns. Thanks to its adaptability across various Arabic text genres, involving literary, scientific, and news corpora, this approach proposes a practical framework for linguistic studies. By leveraging big data techniques, it contributes to the development of machine learning applications, artificial intelligence, and annotated linguistic resources. By virtue of its ability to process large-scale datasets through morphological analysis, the system yields valuable insights into computational linguistics, lexicography, and Arabic language education to support ongoing research in these fields.