Pro-active multi-agent recommender system for travelers

This paper presents a multi-agent recommender system in which agents collaborate with each other to facilitate in providing travel recommendations. A set of rules are organized to be used as knowledge-base that enables travel agents to act in an intelligent and pro-active way. The agents are also in charge of building and updating profiles of travelers. History of the travelers´ selections is kept using a multidimensional rating approach. Maximum-likelihood probability and multi-attribute theory are used to get the recommendation ratings and store them for future recommendations. The effectiveness of the system was evaluated by a diverse set of users with a variety of scenarios. The overall results indicate that the system is capable of pro-actively providing the required types of services with a good degree of accuracy.