Pipelined Filter Combination in Product Personalization

In many cases, standard of-the-shelf products—i.e., products without individual customization— do not meet all customer requirements. Therefore, for instance, the automotive industry produces vehicles with personalized characteristics. Can this customization be further improved w.r.t. production time, especially in B2P relations? Today, on-line ordering systems already store user profiles that can help to match customer requirements. Another approach is mass customization where users can select a product with discrete options (e.g., different front panel color of a stereo). In this paper, we want to go further and allow more than just discrete choices. However, users are not willing to spend hours in front of computers in order to tailor a product—even when it will then fully comply with their wishes. Obviously, we need a mechanism for a computer-supported production of user recommendations. The idea is to let the computer do most of the work in personalizing the product, based on a sophisticated profile and filtering techniques.