A User-centered Approach to Define a Content-based Recipe Similarity Measure
Keywords
Loading...
Authors
Issue Date
2010-08-17
Language
en
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In general, people feel the need to vary in their evening meals. However, especially during busy
week days, when they get back from work, people face a number of barriers to select and prepare a
variety of meals. For people in such a situation, it would be nice to get a recommendation for a meal
that is different from the ones they normally eat, but still fits their regular eating pattern. This thesis
will focus on the identification, application and evaluation of methods to assist people to vary their
food intake by suggesting recipes. These recipe suggestions could be variations to familiar recipes,
which would lead to a more varied food intake without changing eating patterns. To search for such
a variation, we present a method to calculate the similarity between recipes. First, we investigated
which recipe characteristics identify this similarity. Next, we showed how these characteristics can be
extracted from a recipe text and how we used them in a similarity measure. Further, we compared our
similarity measure to the commonly used tf-idf method. The results showed that our measure strongly
outperforms the tf-idf method. Finally, our similarity measure was used for recommending recipes to
people at home. The recommendations were highly appreciated, while several improvements could
still be implemented relatively easy. We will discuss some ways on how to do this.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen