Improvement of viewers´ preference inference by applying the peak-end rule

In this paper, we present a novel scheme to infer viewers´ preferences. The proposed scheme includes the mathematical modeling for pattern analysis of TV viewing behavior based on the peak-end rule in the field of the psychology of consumer behavior. In order to reflect the intensity of interest and the experience time, the experience function is defined and the time-weight average technique is used. From the experimental results, we conclude that the proposed scheme has improved the precision of predicted preference compared with the conventional Bayesian analysis.