Effects and challenges of using a nutrition assistance system: results of a long-term mixed-method study

Abstract

Healthy nutrition contributes to preventing non-communicable and diet-related diseases. Recommender systems, as an integral part of mHealth technologies, address this task by supporting users with healthy food recommendations. However, knowledge about the effects of the long-term provision of health-aware recommendations in real-life situations is limited. This study investigates the impact of a mobile, personalized recommender system named Nutrilize. Our system offers automated personalized visual feedback and recommendations based on individual dietary behaviour, phenotype, and preferences. By using quantitative and qualitative measures of 34 participants during a study of 2–3 months, we provide a deeper understanding of how our nutrition application affects the users’ physique, nutrition behaviour, system interactions and system perception. Our results show that Nutrilize positively affects nutritional behaviour (conditional $$R^2=.342$$ R 2 = . 342 ) measured by the optimal intake of each nutrient. The analysis of different application features shows that reflective visual feedback has a more substantial impact on healthy behaviour than the recommender (conditional $$R^2=.354$$ R 2 = . 354 ). We further identify system limitations influencing this result, such as a lack of diversity, mistrust in healthiness and personalization, real-life contexts, and personal user characteristics with a qualitative analysis of semi-structured in-depth interviews. Finally, we discuss general knowledge acquired on the design of personalized mobile nutrition recommendations by identifying important factors, such as the users’ acceptance of the recommender’s taste, health, and personalization.

Intra-list similarity and human diversity perceptions of recommendations: the details matter

“…Therefore, we are able to compare, across multiple participants, if the order of items, on average, makes a difference on the human perception. -Phase 2: In Phase 2, the ILS metric was based on meta-data: we used genre overlap for the movie domain, as done, e.g., by Vargas et al (2014), and cosine similarity of the Latent Dirichlet Allocation (LDA) embeddings of ingredients and cooking directions as done by Hauptmann et al (2021). We recall that we purposely select different types of meta-data for each domain.…”

“…Ziegler et al (2005) used Amazon's book taxonomy to diversify the results of their recommender. Later works relied on various other types of (meta-)data, for example, movie genres (Vargas et al 2012), food ingredients (Hauptmann et al 2021), artist similarity based on social tags (Jannach et al 2017), or latent topic models from the users' interactions (Shi et al 2012).…”

CarpeDiem: Exploring health pillars to define a holistic recommender system to accomplish a sutainable change in lifestlye habits

“…In this sense, a simplified gamification system with an associated score could help to improve nutrition tracking, as well as taking into considerations limiting factors such as stress, tracking effort, social situations, or exceptional events (e.g. holiday/birthday) [34]. The holistic recommender algorithm in global should also be adaptive, updating the multi-pillar healthy model while new data from the population is obtained.…”

Non-binary Evaluation of Next-basket Food Recommendation

“…In food recommender systems, most performance evaluations have been done almost exclusively in offline experiments [50]. Online evaluations and user studies in the food recommendation research have been conducted in the past few years [31], mostly in the top-n cooking recipe recommendation domain [36,[51][52][53][54]. While most food and recipe recommendation studies have been conducted with study participants in short single experimental sessions or through online crowdsourcing platforms [52,53,55], some studies, especially on the health-aware recipe recommendation, have employed a more rigorous controlled experiment design which took place over several weeks [54,56].…”