Erna Daniati, Aji Prasetya Wibawa, Wahyu Sakti Gunawan Irianto
This research investigates the extraction of narrative events from Andersen’s fairy tales by employing a hybrid BERT-LSTM model to enhance the understanding and analysis of classical literary works through natural language processing (NLP). The model merges the robust language representation capabilities of BERT with the sequential processing strengths of LSTM to capture both the semantic and contextual subtleties of narrative events throughout Andersen’s tales. By leveraging this hybrid approach, the study effectively identifies key narrative components such as character actions, plot developments, and crucial events. The model’s performance was rigorously assessed, achieving a precision of 0.88, a recall of 0.82, and an F1-score of 0.75, demonstrating effective accuracy and a solid balance between detecting true positives and reducing false negatives. The results contribute to computational literary analysis, providing a dependable method for analyzing complex narrative structures in fairy tales and similar texts. This method not only enriches literary research with quantitative insights but also has potential applications in automated story generation, educational tools, and digital humanities. Future research will aim to improve recall rates and further refine the model for broader application across various literary genres. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.
Universitas Negeri Malang, Malang, Indonesia
