Communications of the ACM
The Online Dating Engine That Assesses Your Taste in the Opposite Sex (and Whether They Find You Attractive)
A new dating recommendation engine analyzes the number of replies a user receives on a dating site, and uses this to rate a person's relative attractiveness.
credit: looooker.com
University of Iowa computer scientist Kang Zhao and colleagues have created a dating recommendation engine that suggests potential dates based not only on mutual interests, but also on a person's likelihood to reply to initial contact.
Many dating websites make recommendations based on the preferences of other users who have shown an interest in similar people, but fail to consider a user's attractiveness to potential dates based on the number of replies they receive. Zhao's recommendation engine analyzes the number of replies a user receives and uses this to rate a person's relative attractiveness.
The researchers conducted tests with anonymized data from a dating website with 47,000 users over 196 days, using the first 98 days of data as a training set to identify the tastes and attractiveness of individual users. The new engine's recommendations were compared with other methods, such as using taste only, on the basis of the number of potential dates recommended as well as how often contacts were reciprocated.
The team notes that beyond dating, their engine could be used in job application networks, which are similar to heterosexual dating networks in that both are reciprocal bipartite networks that contain two types of nodes, with links created between different node types.
From Technology Review
View Full Article
Abstracts Copyright © 2013 Information Inc., Bethesda, Maryland, USA
No entries found