Affective Computational Advertising Based on Perceptual Metrics
We present \textbf{ACAD}, an \textbf{a}ffective \textbf{c}omputational \textbf{ad}vertising framework expressly derived from perceptual metrics. Different from advertising methods which either ignore the emotional nature of (most) programs and ads, or are based on axiomatic rules, the ACAD formulation incorporates findings from a user study examining the effect of within-program ad placements on ad perception. A linear program formulation seeking to achieve (a) \emph{genuine} ad assessments and (b) \emph{maximal} ad recall is then proposed. Effectiveness of the ACAD framework is confirmed via a validational user study, where ACAD-induced ad placements are found to be optimal with respect to objectives (a) and (b) against competing approaches.