Towards a Web-based Multi-Criteria Decision Support System for the Layered Evaluation of Interactive Adaptive Systems
Résumé
The choice of suitable evaluation methods for the layered evaluation of Interactive Adaptive Systems (IAS) needs the consideration of different factors, leading to a multi-criteria decision analysis problem. This paper proposes the first step toward a Web-based Multi-Criteria Decision Support System (MCDSS). Our proposal is based on a recent multi-criteria decision method called ELECTRE TRI-B-H (Elimination and Choice Translating Reality) to guide the layered evaluation of IAS. The goal is to support the choice of appropriate evaluation methods for individual layers taking into account constraints of the layered evaluation and the individual layers. The appropriateness of each evaluation method is analyzed for the layered evaluation in general and each layer in particular. The use of ELECTRE TRI-B-H method allows decomposing a decision problem into intermediate sub-problems through a hierarchy model and sorting alternative evaluation methods at different levels of the hierarchy. A case study of an adaptive hypermedia system is presented; in this study, the ELECTRE TRI-B-H method has been applied to select the most suitable evaluation methods for each of its layers taking into account the layered evaluation context. The promising results show the feasibility of the proposed approach, leading to various research perspectives