The researchers of this study described a recommender system they implemented and performed a quantitative comparison of two collaborative filtering (CF) and two global algorithms.Results showed that collaborative filtering recommenders significantly outperform global algorithms used by dating sites.
Unlike previous studies that relied solely on self-report data, this study establishes ground truth for 80 online daters’ height, weight and age, and compares ground truth data to the information provided in online dating profiles.
The benefits of the proposed methodology with respect to traditional matchmaking baseline systems are shown by an extensive evaluation carried out using data gathered from a real online dating service.
This analysis also provides deep insights into the aspects of matchmaking that are important for presenting highly relevant matches.
Dating agency members (second group) completed similar questionnaires examining partner preferences.
Again, here preferences were similar across the sexes, although men preferred a submissive and introverted woman and stressed the importance of physical appearance in a mate.