نوع مقاله : مقاله پژوهشی
نویسنده
پژوهشگاه علوم و فناوری اطلاعات ایران (ایرانداک)
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسنده [English]
Topic extraction and classification of Hafez's poetry has long been a focus of scholarly attention, with existing studies primarily adopting either historical periodization approaches or relying on expert interpretations from Persian literature specialists. This study presents a novel computational methodology for identifying and categorizing topics in Hafez's Divan using artificial intelligence and text mining techniques, completely independent of human expert judgment.
Following text preprocessing and feature extraction, our analysis revealed eight principal topics within the Divan. These topics were automatically labeled based on their most characteristic vocabulary using AI assistance: 1) Anguish and Ardor, 2) Panegyric and Praise, 3) Separation and Reunion, 4) Beauty and Allure, 5) Revelry and Intoxication, 6) The Mystery of Love, 7) Mysticism and Spirituality, and 8) Bohemianism and Libertarianism.
Each poem was systematically classified according to its strongest topic association, with threshold-based analysis employed to determine significant thematic connections. Our completely automated, machine learning-driven approach to lexical pattern analysis in Hafez's work demonstrates a reproducible methodology applicable to other literary corpora.
کلیدواژهها [English]
)