E-Commerce site evaluations, comparisons and recommendations within a possibilistic framework

International audienceFaced with the important amount of information and the multitude of potential choices on the Web, indecisive Internet consumers tend to turn to the opinions and experience of the other customers to make their choices. This behaviour is at the origin of the concept of e-recommendation. This paper proposes formalizing e-recommendation as a process of multi-criteria choice. The uncertainty related to the imprecision, the dispersion and the divergence of the opinions of the customers is modelled by a possibilistic representation. An aggregation based on the Choquet integral is used to take interactions between criteria into account. This paper is more specifically concerned with the comparison of e-commerce websites in an uncertain universe and takes up the multi modality aspects of the evaluations