Recommender systems support internet users in finding and identifying products and services suiting their individual demands (e.g.: digital cameras, financial services). Although recommender systems already offer mechanisms which alleviate the comparison of different suitable products (e.g. product lists, comparison pages) users usually have difficulties in decision making and finding the optimal option. Persuasive mechanisms can be used in such situations for underlining product differences and reinforcing confidence in the users' own decision. This leads to an increase of the trust level and supports the decision making process. Especially theories concerning user behaviour in buying situations constitute great potential for persuasion in recommender systems. It has been shown that the user's perception of the value of a certain product is highly influenced by the context (i.e. the set of presented products in the choice set). Well-known context effects are the asymmetric dominance effect, the attraction effect, or the compromise effect. This paper presents a multi-preferential and multi-alternative model (i.e. more than two product attributes and more than two products are supported) for calculating dominance values of items in choice sets and thus offers the possibility of determining the best recommendation set in a given choice situation. The performance of the model is shown by the application on empirical data (choice sets) gained by a previously conducted user study.
展开▼
