Boosting Healthiness Exposure in Category-constrained Meal Recommendation Using Nutritional Standards

Ming Li,Lin Li,Xiaohui Tao,Zhongwei Xie,Qing Xie,Jingling Yuan
DOI: https://doi.org/10.1145/3643859
IF: 5
2024-02-05
ACM Transactions on Intelligent Systems and Technology
Abstract:Food computing, as a newly emerging topic, is closely linked to human life through computational methodologies. Meal recommendation, a food-related study about human health, aims to provide users a meal with courses constrained from specific categories (e.g., appetizers, main dishes) that can be enjoyed as a service. Historical interaction data, as important user information, is often used by existing models to learn user preferences. However, if user’s preferences favor less healthy meals, the model will follow that preference and make similar recommendations, potentially negatively impacting the user’s long-term health. This emphasizes the necessity for health-oriented and responsible meal recommendation systems. In this paper, we propose a healthiness-aware and category-wise meal recommendation model called CateRec, which boosts healthiness exposure by using nutritional standards as knowledge to guide the model training. Two fundamental questions are raised and answered: 1) How to evaluate the healthiness of meals? Two well-known nutritional standards from the World Health Organisation and the United Kingdom Food Standards Agency are used to calculate the healthiness score of the meal. 2) How to health-orientedly guide the model training? We construct category-wise personalization partial rankings and category-wise healthiness partial rankings, and theoretically analyze that they meet the necessary properties and assumptions required to be trained by the maximum posterior estimator under Bayesian probability. The data analysis confirms the existence of user preferences leaning towards less healthy meals in two public datasets. A comprehensive experiment demonstrates that our CateRec effectively boosts healthiness exposure in terms of mean healthiness score and ranking exposure, while being comparable to the state-of-the-art model in terms of recommendation accuracy.
computer science, information systems, artificial intelligence
What problem does this paper attempt to address?
### Problems Addressed by the Paper The paper aims to address the issue of how to increase the exposure of healthy meals in category-constrained food recommendation systems. Specifically, existing food recommendation systems typically rely on users' historical interaction data to learn their preferences and generate new recommendations accordingly. However, if users' preferences lean towards unhealthy meals, these systems may continue to recommend similar unhealthy meals, negatively impacting users' long-term health. Therefore, this paper proposes a health-aware, category-aware food recommendation model (referred to as CateRec), which uses nutritional standards as knowledge to guide model training, thereby increasing the exposure of healthy meals. ### Main Contributions 1. **Highlighting the Unhealthy Recommendation Phenomenon**: The paper points out the prevalent issue of unhealthy recommendations in existing food recommendation systems and proposes a health-aware, category-aware food recommendation model, CateRec, which uses nutritional standards to guide model training and increase the exposure of healthy meals. 2. **Theoretical Analysis**: Provides theoretical analysis demonstrating that the proposed category-aware health partial ranking training method meets the necessary properties and assumptions required by common training paradigms, namely the maximum a posteriori estimator under Bayesian probability. 3. **Experimental Validation**: Experimental results on 2 public datasets show that CateRec effectively increases the exposure of healthy meals in terms of average health score and ranking exposure, while maintaining comparable recommendation accuracy to state-of-the-art models. ### Method Overview - **Health Assessment**: Uses nutritional standards from the World Health Organization (WHO) and the UK's Food Standards Agency (FSA) to calculate the health scores of meals. - **Model Training**: Constructs category-aware personalized partial ranking and health partial ranking, demonstrates their rationality through theoretical analysis, and uses these rankings to guide model training during the training process. ### Conclusion By proposing the CateRec model, the paper effectively addresses the issue of insufficient exposure of healthy meals in food recommendation systems, providing new ideas and methods for achieving health-oriented and responsible food recommendation systems.