Clustering analysis of metered Arabic poetry compositions

This article focuses on identifying the defining features of a metered Arabic poetic structure. In addition to its research and educational advantages, emphasizing the comprehensibility of the structure, even without language proficiency, enables individuals to grasp its core essence. This understanding opens up opportunities for exploring comparative possibilities and establishing connections with poetic structures in different
languages.
To achieve this goal, this article presents a novel clustering approach based on the six following poem metrics (Diversity, Evenness, Variability, Repetition, Pattern Potential and Pattern Usage) and utilizing two distinct algorithms. We discovered that the most significant poem metrics are Repetition, Variability and Pattern Usage and two attributes are sufficient to characterize the poetry instead of six.