You Are What You Eat: Learning User Tastes for Rating Prediction

Harvey, Morgan; Ludwig, Bernd; Elsweiler, David

doi:10.1007/978-3-319-02432-5_19

Cited by 61 publications

(63 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Work in food recommender systems has typically focused on rating prediction, utilising recipe content [14,33] and contextual information [17] to minimise prediction error. Harvey et al [17] showed that one important contextual factor for recipe recommendation is the users themselves: a small group explicitly preferred healthier food, while the majority tended to prefer less healthy alternatives. Recent work has tried to incorporate healthiness into the recommendation process by substituting ingredients [33], incorporating calorie counts [16], and generating food plans [12].…”

Section: Related Workmentioning

confidence: 99%

“…One approach to optimising this trade-o would be to substitute meals that would typically be recommended to users (as in [14,17]) with similar but healthier dishes. For this strategy to be successful, however, a number of prerequisites need to be ful lled: 1) recipes need to exist that are su ciently similar in style and content, but di erent in health properties.…”

Section: Related Workmentioning

confidence: 99%

“…People search for recipes in a variety of ways for many di erent purposes [8,42] and a relatively large proportion of web search queries are related to food or lead to the visit of a food-related website [39]. As such, systems which provide access to online recipes or make personalised recommendations have been touted as a means to help people nourish themselves more healthily [14,17]. Nevertheless, despite o ering access to healthy content [34], analyses of the systems being used in practice indicate that they tend to promote unhealthy meals [35].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Exploiting Food Choice Biases for Healthier Recipe Recommendation

Elsweiler

Trattner

Harvey

2017

Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval

Self Cite

138

View full text Add to dashboard Cite

By incorporating healthiness into the food recommendation / ranking process we have the potential to improve the eating habits of a growing number of people who use the Internet as a source of food inspiration. In this paper, using insights gained from various data sources, we explore the feasibility of substituting meals that would typically be recommended to users with similar, healthier dishes. First, by analysing a recipe collection sourced from Allrecipes.com, we quantify the potential for nding replacement recipes, which are comparable but have di erent nutritional characteristics and are nevertheless highly rated by users. Building on this, we present two controlled user studies (n=107, n=111) investigating how people perceive and select recipes. We show participants are unable to reliably identify which recipe contains most fat due to their answers being biased by lack of information, misleading cues and limited nutritional knowledge on their part. By applying machine learning techniques to predict the preferred recipes, good performance can be achieved using low-level image features and recipe meta-data as predictors. Despite not being able to consciously determine which of two recipes contains most fat, on average, participants select the recipe with the most fat as their preference. The importance of image features reveals that recipe choices are often visually driven. A nal user study (n=138) investigates to what extent the predictive models can be used to select recipe replacements such that users can be "nudged" towards choosing healthier recipes. Our ndings have important implications for online food systems. KEYWORDSFood RecSys; human decision making; behavioural change; information behaviour ACM Reference format:

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Exploiting Food Choice Biases for Healthier Recipe Recommendation

Elsweiler

Trattner

Harvey

2017

Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval

Self Cite

138

View full text Add to dashboard Cite

show abstract

“…With this aim, the recipe recommendation system proposed in (Freyne and Berkovsky, 2010) makes tailored recommendations of healthy recipes; in (Ueda et al, 2011) authors created a personalized recipe recommendation method that is based on the user's food preferences from his/her recipe browsing and cooking history; in (Harvey et al, 2013) authors presented a system that learns user tastes for improving the rating prediction and make better recommendations of recipes; and in (Elsweiler and Harvey, 2015) the system recommends recipes that users will like and that fit into a balanced diet. In the area of persuasion technologies, the Portia system (Mazzotta et al, 2007) used relevant arguments, both rational and emotional, to persuade people to change their eating habits.…”

Section: Related Workmentioning

confidence: 99%

Bargaining agents based system for automatic classification of potential allergens in recipes

Alemany-Bordera

Barberá

Pérez-López

et al. 2016

ADCAIJ

View full text Add to dashboard Cite

Recommendation system; Food allergy; Multi-agent systemThe automatic recipe recommendation which take into account the dietary restrictions of users (such as allergies or intolerances) is a complex and open problem. Some of the limitations of the problem is the lack of food databases correctly labeled with its potential allergens and non-unification of this information by companies in the food sector. In the absence of an appropriate solution, people affected by food restrictions cannot use recommender systems, because this recommend them inappropriate recipes. In order to resolve this situation, in this article we propose a solution based on a collaborative multi-agent system, using negotiation and machine learning techniques, is able to detect and label potential allergens in recipes. The proposed system is being employed in receteame.com, a recipe recommendation system which includes persuasive technologies, which are interactive technologies aimed at changing users' attitudes or behaviors through persuasion and social influence, and social information to improve the recommendations. MotivationWhen the doctor diagnoses a patient with an allergy, the set of substances that produce extreme sensitivity in the patient's immune system should be avoided. In the case of food allergies, the difficulty is greater because avoiding a food is not an easy task when many of them are composed by others (e.g. mayonnaise, is composed of oil, egg, vinegar, etc.). This is an awkward situation of high impact because it involves nutrition, a necessary and daily task in the lives of people activity. This increasingly affects to more individuals in our society (up to 8% in children and 2% in adults) 1 . Moreover, the problem of food allergies is not resolved by simply avoiding certain foods, because the lack of nutrients they provide must be compensated with other foods.People affected by food allergies are forced to become expert nutritionists to maintain a healthy life, free from allergens that they cannot tolerate. Currently, Internet is the most popular way of obtaining information about allergies. On the Internet, for instance, the World Allergy Organization (WAO) regulates and offers the terminology used to characterize allergies information. However, the information is difficult to understand because of its complexity and quantity. Thus, traditional search and navigation activities are being combined or even replaced by direct interactions between users in the form of recommendations, advice and warnings; 2 out of 3 take into account the recommendations of other users to make decisions (about products, treatments, entertainment, etc.); and of these, 69% gives a lot or some credibility to what their friends or acquaintances say on social networks. 1 WAO World Allergy Organization, Food allergy statistics: http://www.worldallergy.org/public/allergic_diseases_center/foodallergy/

show abstract

“…Retrieving and structuring the information available in a food recipe document allows for finetuned search, and to improve food recommendation systems [2]. Consequently, it serves as the basis of more accurate nutritional calculations, bringing further clarity on the nutritional content of the meal and overall effect on health and well-being [3]. Available related studies have focused on extracting a broad amount of information from recipes, with the goal of abstracting the recipe text into different types of representations [4] [5], but few studies were found that attempt to extract this information at a more granular level, not strictly with the goal of representing the recipe but, instead, to serve as the foundation to extract further nutritional-related information.…”

Section: Introductionmentioning

confidence: 99%

Information Extraction from Unstructured Recipe Data

Silva

Ribeiro

Ferreira

2019

Proceedings of the 2019 5th International Conference on Computer and Technology Applications

View full text Add to dashboard Cite

Online food recipes are an important source of information for many individuals, who use these to learn how to cook new dishes and choose their meals. However, these often lack structured information, useful to improve search and recommendation systems of food recipe websites, as well as calculate accurate nutritional information, which brings additional value to users. To solve this problem, FRIES was developed. FRIES automatically extracts the names, quantities, units and cooking methods for each ingredient in a recipe. The system uses mainly rule-based methods and achieves an average F-measure of 0.89 for the extraction of the cooking methods present in a recipe and an average F-measure of 0.83 for the extraction of associations linking cooking methods to ingredients. FRIES' results show that it can accurately and automatically extract information from cooking recipes. This information can be used to estimate the nutritional information of food recipes and support recommendation systems.

show abstract

You Are What You Eat: Learning User Tastes for Rating Prediction

Cited by 61 publications

References 8 publications

Exploiting Food Choice Biases for Healthier Recipe Recommendation

Exploiting Food Choice Biases for Healthier Recipe Recommendation

Bargaining agents based system for automatic classification of potential allergens in recipes

Information Extraction from Unstructured Recipe Data

Contact Info

Product

Resources

About