Abstract: A multimode video merchandiser system utilizes two levels of inductive learning to derive rules for selecting the sequence in which images of products stored on a videodisc are presented on a video monitor to a user. The first level of inductive learning generates rules from market survey based, consumer profile attributes assigned to items selected by previous users to determine the profile of the consumer most likely to be using the system at any given time, and to present the items in a sequence most likely to appeal to such a user. The second level of inductive learning utilizes a set of product characteristic attributes assigned to items selected by the current user to determine that user's preferences, and to modify the sequence of presentation to display first those items possessing the preferred characteristics.