قياس التغيرات الزمنية في المصطلحات التاريخية عبر "Text Mining".

Abstract

Over the past two decades, Text Mining technologies have revolutionized the study of language and history, enabling researchers to analyze large-scale historical corpora with greater accuracy and objectivity than traditional methods. This study employs Natural Language Processing (NLP) and text mining techniques to identify and analyze changes in historical terminology across multiple time periods using large, chronologically organized digital historical corpora. The significance of this research lies in its ability to trace the emergence, evolution, and semantic transformation of historical terms, while examining their relationship with the social, political, and cultural developments of the societies in which they were used. The study applies computational methods such as Term Frequency–Inverse Document Frequency (TF–IDF) to identify significant terms across different historical periods, alongside Topic Modeling techniques to uncover the semantic patterns characterizing each era. Recent findings (2025) from studies on European historical texts using Explainable Machine Learning (XML) demonstrate that linguistic patterns associated with specific historical periods can be predicted with an accuracy of up to 76.7% in distinguishing between different centuries. Feature importance analysis further revealed clear patterns of linguistic evolution over time. These findings highlight the effectiveness of digital text mining in historical research, not only for tracking terminological change but also for deepening our understanding of the relationship between language and the socio-political contexts in which historical terminology developed. Consequently, this approach opens new avenues for research in historical linguistics and the digital humanities.