[Paper Review] Food Recommender Systems: Important Contributions, Challenges and Future Research Directions

Christoph Trattner, David Elsweiler|arXiv (Cornell University)|Nov 7, 2017

Recommender Systems and Techniques38 references67 citations

TL;DR

The paper surveys the state-of-the-art in food recommender systems, covering recipe, meal-plan, grocery, and menu recommendations, reviews approaches (content-based, collaborative filtering, hybrid), and discusses health, context, group settings, and implementation resources with open challenges and future directions.

ABSTRACT

The recommendation of food items is important for many reasons. Attaining cooking inspiration via digital sources is becoming evermore popular; as are systems, which recommend other types of food, such as meals in restaurants or products in supermarkets. Researchers have been studying these kinds of systems for many years, suggesting not only that can they be a means to help people find food they might want to eat, but also help them nourish themselves more healthily. This paper provides a summary of the state-of-the-art of so-called food recommender systems, highlighting both seminal and most recent approaches to the problem, as well as important specializations, such as food recommendation systems for groups of users or systems which promote healthy eating. We moreover discuss the diverse challenges involved in designing recsys for food, summarise the lessons learned from past research and outline what we believe to be important future directions and open questions for the field. In providing these contributions we hope to provide a useful resource for researchers and practitioners alike.

Motivation & Objective

• Summarize the state-of-the-art in food recommender systems across recipes, meal plans, groceries, and menus.
• Identify seminal and recent approaches and their performance in food recommendation.
• Discuss domain-specific challenges (health, context, group recommendations, dietary constraints) and lessons learned.
• Outline open questions and future directions for researchers and practitioners.

Proposed method

• Review of literature and tabular synthesis of 25 key papers with attributes (algorithm, item type, feedback, context, dietary constraints, target, dataset).
• Categorization of approaches into content-based (CB), collaborative filtering (CF), and hybrid methods.
• Analysis of context-aware, group-based, and health-aware recommender strategies.
• Discussion of implementation resources including datasets, nutrition resources, and frameworks.

Experimental results

Research questions

• RQ1 What algorithms and data representations have been most effective for food item recommendations (recipes, meals, groceries, menus)?
• RQ2 How have CB, CF, and hybrid methods been adapted to food-specific constraints (allergies, dietary patterns, health goals)?
• RQ3 What is the role of context (time, location, gender, availability) and group settings in food recommendations, and how do health considerations interact with accuracy?
• RQ4 What datasets and resources exist to develop and evaluate food recommender systems, and what are the main open challenges and future directions?

Key findings

• Content-based methods leverage ingredients and nutritional aspects to tailor recipes to user tastes.
• Collaborative filtering approaches (notably LDA and WRMF) often outperform basics like basic CF or MostPopular on recipe data.
• Hybrid approaches can balance accuracy and health considerations, but post-filtering for health may reduce recommendation quality; optimal trade-offs remain an open area.
• Contextual factors (time, gender, location, availability) influence ratings and bookmarks, indicating the need for context-aware models.
• Group-based strategies improve group satisfaction but personalization per individual remains challenging.
• Health-aware methods show potential but achieving high nutritional quality without sacrificing accuracy is difficult; long-horizon planning (daily/weekly) could help.