Improving Indonesian Recipe Named Entity Recognition via Weighted Ensemble and Semi-Supervised Learning

Dody Haryadi (School of Data Science, Mathematics and Informatics, IPB University, Bogor, Indonesia Informatics Study Program, University Al-Azhar Indonesia, Jakarta)
Imas Sukaesih Sitanggang (School of Data Science, Mathematics and Informatics, IPB University, Bogor)
Sony Hartono Wijaya (School of Data Science, Mathematics and Informatics, IPB University, Bogor)
Budi Setiawan (Department of Community Nutrition, IPB University, Bogor)

Journal of Information Systems Engineering and Business Intelligence
Vol. 12 No. 3 (2026): October

Background: The rapid growth of online culinary platforms has produced a vast amount of unstructured Indonesian recipe texts. Extracting structured information such as ingredients, quantities, and measurement units002C is essential for downstream applications, including nutritional computation and diabetes-oriented dietary recommendation systems. However, Indonesian recipe texts present substantial challenges due to linguistic variability, informal writing styles, and limited availability of annotated corpora.  Objective: This study investigates effective modeling strategies for Indonesian recipe Named Entity Recognition (NER) in low-resource settings and to identify reliable approaches to improve the robustness and generalization of extraction.  Methods: A quantitative experimental approach was employed to evaluate the IndoBERT, BiLSTM-CRF, IndoBERT-CRF, and hybrid IndoBERT-BiLSTM-CRF architectures. A reliability-weighted voting ensemble combining heterogeneous models was proposed, and high-confidence ensemble predictions were used for pseudo-labeling in a semi-supervised setting. Performance was assessed using entity-level precision, recall, and F1-score.  Results: IndoBERT achieved a strong baseline performance (F1-score = 96.61%). IndoBERT-CRF and the hybrid IndoBERT-BiLSTM-CRF model achieved competitive F1-scores (92.99% and 92.82%, respectively) but did not outperform the Transformer-only baseline. The proposed weighted ensemble achieved the best supervised performance (F1-score = 96.79%). Semi-supervised training with ensemble-generated pseudo-labels maintained competitive performance (F1-score = 96.72%) while improving recall and generalization to unseen ingredient expressions.  Conclusion: Increased architectural complexity does not necessarily improve performance on the Indonesian recipe NER task. Reliability-weighted ensemble learning combined with pseudo-labeling provides a stable and effective approach for low-resource culinary NER and supports future nutrition-oriented recipe processing.  Keywords: Indonesian Recipe Text; Pseudo-Labeling; Named Entity Recognition; Semi-Supervised Learning; Weighted Ensemble