EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Leakage-Aware Machine Learning and Deep Learning Benchmarking of Food Antioxidant Capacity Prediction on the Antioxidant Food Table

Erkan Caner Ozkat iD

Antioxidants · 11.09.2026 · 15 (9) · 1157

DOI: 10.3390/antiox15091157 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Antioxidants, online erschienen am 11.09.2026. Verfasst von Erkan Caner Ozkat. Beteiligte Einrichtung: Department of Mechanical Engineering, Faculty of Engineering and Architecture, Recep Tayyip Erdogan University, 53100 Rize, Türkiye. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 67 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Recep Tayyip Erdoğan University, Recep Tayyip Erdogan University Development Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The Antioxidant Food Table is the largest open collection of measured food antioxidant capacity. It covers 3139 products assayed by the ferric reducing ability of plasma (FRAP) method. The table has served mainly as a dietary lookup source and has never been machine-readable or benchmarked. Here it was extracted into a validated open dataset (3135 records, 99.9%). A leakage-aware benchmark of antioxidant capacity prediction from product description and category was then constructed. Eighteen predictors, from naïve baselines to deep networks and fusions, were evaluated under two partitioning regimes with the same five seeds and permutation controls. Since 39.4% of records share a product name, the conventional random split rewards memorization. A learning-free duplicate lookup explained 45% of the apparent k-nearest-neighbor advantage over the category median. In grouped evaluation, ridge regression on term frequency–inverse document frequency (TF–IDF) features (R2=0.674) outperformed both deep networks. Pretrained word vectors did not close this gap. An equal-weight fusion of all eight models performed best (R2=0.684) with 2.6-fold lower variability. Protocol and representation, rather than architecture, dominated the outcome on this benchmark. The dataset, code, and predictions are released openly.

Schlagwörter

Stichwörter der Autor:innen

antioxidant capacitydata leakagedeep learningfood qualitymachine learning

Autor:innen und Institutionen

  • Department of Mechanical Engineering, Faculty of Engineering and Architecture, Recep Tayyip Erdogan University, 53100 Rize, Türkiye.

ORCID-Profile: Erkan Caner Ozkat

Förderung

  • Recep Tayyip Erdoğan University
  • Recep Tayyip Erdogan University Development Foundation

Zitieren

APA

Ozkat, E. C. (2026). Leakage-Aware Machine Learning and Deep Learning Benchmarking of Food Antioxidant Capacity Prediction on the Antioxidant Food Table. Antioxidants, 15(9), 1157. https://doi.org/10.3390/antiox15091157

Vancouver

Ozkat EC. Leakage-Aware Machine Learning and Deep Learning Benchmarking of Food Antioxidant Capacity Prediction on the Antioxidant Food Table. Antioxidants. 2026;15(9):1157. doi:10.3390/antiox15091157

BibTeX
@article{Ozkat2026,
  author = {Ozkat, Erkan Caner},
  title = {Leakage-Aware Machine Learning and Deep Learning Benchmarking of Food Antioxidant Capacity Prediction on the Antioxidant Food Table},
  journal = {Antioxidants},
  year = {2026},
  volume = {15},
  number = {9},
  pages = {1157},
  doi = {10.3390/antiox15091157}
}
Kostenlos starten →