The Case-Based Matchmaker

Matchmaking of human beings is recognized as a difficult task requiring great insight and sensitivity. In this paper, we present a case-based model that enlarges the number of relevant matches for any applicant. This model has been implemented in a working system for matches among either religious or traditional Jews in Israel. One implemented idea is the regular case-based retrieval, i.e. retrieving proposals similar to previous successful proposals. In addition, we implement a more complex case-based retrieval, which retrieves proposals in a few case-based steps. The model has been implemented in a working system intended for use in real life. Examples of the experiment we carried out show that casebased retrieval helps the majority of people by proposing new proposals in addition to those proposed in the regular kinds of retrieval. We think that these case-based retrieval methods, in principle, can be generalized to a certain extent to other matching tasks. Another future idea is to evaluate the potential of the case-based matchmaker as an intelligent support system for matchmakers.