Model Reduction in User Adaptive Emotion-based Selection Systems

A simple way of implementing user adaptivity in emotion-based selection systems is the adaptation of behaviour-based control structures [6], [7]. In this case the user adaptation itself is handled as a kind of adaptive fusion of existing (collected off-line) human opinions (emotional models). Beyond the difficulties of collecting emotional models related to a set of objects from numerous individuals, these systems are also faced with the problem of economical model storage. For tackling the problem of model storage, the adaptation of HOSVD (High Order Singular Value Decomposition) based model reduction [10], [11] is proposed, as a kind of model pre-processing and handling method.