Akademik Veri Yönetim Sistemi
9th ICONST 2025 - Kotor International Conferences on Science and Technology , Kotor, Karadağ, 3 - 05 Eylül 2025, ss.26, (Özet Bildiri)
Abstract: The role of Artificial Intelligence (AI) is increasing every day. There are many AI based methods used for different areas. This study focuses on the analysis of emotional and conceptual structures of literary texts by means of AI based methods. By applying Natural Linguistic Processing (NLP) techniques on literary text written in Turkish language, sensual density and conceptual context are inferred. This study makes it possible to combine the expertise of literary analysis with technical knowledge of machine learning and NLP and signal processing and data visualization. In the study, first, poems and short stories belonging to distinct era have been chosen and converted to the digital format. Additionally, besides the grammatical pre-processing stages on the texts, BERT based models have been used for the classification of emotion. Moreover, symantic proximity analysis has been made with word vectorization methods such as word2vec and fastText. The obtained emotion data has been modeled in time series and emotional transition. The emotion data have been visualized by color coded heat maps and graphics and a prototype an interactive interface has been developed. The results show that AI based analysis permits deeper and objective examination of literary texts. As a result, the study has the potential to contribute to literature education. The study brings a new perspective on digital human science with interdisciplinary methods and presents significant evidence of the use of AI techniques in the analysis of literary texts. Keywords: Literary texts, NLP, AI, Emotion analysis