Predicting Nutrient Density in Foods Using Machine Learning Models: A Comparative Study

Changhe Yang

2024

Abstract

Fine-tuning and thus accurately estimating nutrition density in foods is useful in optimizing diets and improving health standards. Several challenges have been observed with the traditional methods for nutrient evaluation. Most of these challenges can be trimmed down by adopting the use of Machine Learning (ML) models, which possess better capabilities of giving efficient and accurate assessments. To this end, different regression models were applied to estimate nutrient density, namely Linear Regression, Ridge Regression, Decision Trees, and Random Forests. The used set had 2397 food items for which 33 nutrients had been deemed relevant. The missing values in the chosen dataset were addressed before model training through imputation and normalization for better data quality. The models were trained and evaluated using separate training and test sets, with performance indicators such as Mean Absolute Error (MAE) and R-squared (R²) used to measure their accuracy. Results showed that linear models, such as Linear Regression and Ridge Regression, achieved the best accuracy, with an R² of 0.999, while tree-based models exhibited overfitting tendencies, resulting in lower predictive performance on unseen data. These findings demonstrate the effectiveness of machine learning in predicting nutrition density, significantly improving the precision of dietary recommendations.

Paper Citation

in Harvard Style

Yang C. (2024). Predicting Nutrient Density in Foods Using Machine Learning Models: A Comparative Study. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 64-69. DOI: 10.5220/0013487500004619

in Bibtex Style

@conference{daml24,
author={Changhe Yang},
title={Predicting Nutrient Density in Foods Using Machine Learning Models: A Comparative Study},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={64-69},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013487500004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Predicting Nutrient Density in Foods Using Machine Learning Models: A Comparative Study
SN - 978-989-758-754-2
AU - Yang C.
PY - 2024
SP - 64
EP - 69
DO - 10.5220/0013487500004619
PB - SciTePress