Ranking preferences deduction based on semantic similarity for the stable marriage problem

Abstract

The stable marriage problem is a well-known problem with many practical applications. Most algorithms to find stable marriages assume that the participants explicitly express a preference ordering. This can be problematic when the number of options is large or has a combinatorial structure. We show, by simply asking the actors (men and women) to fulfill a personal profile with items positioning in a tree-structured semantic network, that it is possible to solve the problem of stable marriages without asking the actors to explicitly operate a ranking over the members of the opposite sex.