تصنيف النصوص العربية باستخدام نظم التنقيب في البيانات مفتوحة المصدر: دراسة تطبيقية باستخدام (WEKA - Orang data mining)

WEKA and Orange Data Mining are considered among the best open-source artificial intelligence applications for Arabic text classification. Both systems allow for building models to classify Arabic texts and each is characterized by a set of features that distinguish it from the other. These features span from the installation process, data preprocessing, model training and testing, to effectively and accurately present the results. Based on text classification using the two systems (Orange Data Mining and WEKA), the results demonstrated that Orange Data Mining significantly outperformed WEKA across all algorithms used to classify texts by specialties, reflecting its high capability in categorizing data into precise classes. Regarding the classification of texts by academic departments, most algorithms achieved similarly excellent performance in both systems, with accuracy rates exceeding 95%. Evaluation and comparison results between WEKA and Orange Data Mining revealed that Orange Data Mining surpassed WEKA in six areas: general requirements, system interface, AI requirements, algorithms, performance speed, and support and maintenance. On the other hand, WEKA outperformed Orange in two areas: result accuracy and reporting/statistics (outputs). Both systems were equal in six other areas: user support, operating requirements, multimedia processin.