Agent technology recommending personalized information and its evaluation

These days, personalized contents recommending services are attracting world wide attention due to the development of broadcasting technology based on WWW and P2P technology. In this paper, we propose a method of implementing our active autonomous agents that discover and recommend contents matched with the user´s preference without the user´s guidance. The agents can learn the user preference by observing the user´s reaction to contents recommended by the agents. We evaluate the convergence and adaptability of our agents´ learning algorithm by simulation with our programmable movie service (PM service), which composes and provides appropriate contents for users.