Towards Health-Aware in Food Recipe Recommendation

The primary focus of my research has been the development and enhancement of Food Recommender Systems (FRS), innovative tools designed to promote healthier eating habits by suggesting personalized food options. As people strive to make healthier dietary choices, FRS assist users in navigating the vast array of food options, reducing the overwhelming amount of information available online and supporting individuals in selecting balanced and nutritious meals. Traditionally, many FRS have based their recommendations solely on a user's past ratings or favorite items, often overlooking key factors such as ingredients and the dynamic nature of user preferences. Additionally, traditional FRS lack transparency, making it unclear to users why a particular recommendation was made. This absence of explainability, coupled with limited consideration for health-related and emotional aspects, can reduce users' trust and engagement with these systems.
In response, this research addresses these limitations by introducing several new methods to improve FRS. I explored approaches to make these systems more adaptable to the changing preferences of individual users and groups, integrating critical aspects such as ingredient information, health goals, and emotional insights. Furthermore, I developed explainable FRS models that reveal the reasoning behind each recommendation, empowering users to understand and trust the system’s suggestions. Other models focused on fairness and health aspects, ensuring that all users have equitable access to health recommendations, regardless of their background or dietary preferences. The research outcomes demonstrate that incorporating these factors enhances both the relevance and accuracy of recommendations. Our models achieved a significant improvement in user satisfaction by enabling individuals to actively balance their health goals with personal taste preferences.
By helping users make healthier dietary choices, these FRS have the potential to positively influence public health outcomes and align with global health goals proposed by the WHO. Additionally, by incorporating fairness and explainability, these systems not only support healthier lifestyles but also promote inclusivity and accountability, empowering users to make better-informed choices that contribute to well-being across diverse communities. Through this research, I aim to create FRS that not only meet dietary needs but also foster healthier, more transparent, and equitable food recommendation experiences for all users.