Application of fuzzy multi attribute decision making analysis to rank web services

In this paper, we employed modified fuzzy multi attribute decision making (FMADM) to rank web services. The modification to the FMADM is that we employed backpropagation trained neural network (BPNN) instead of the analytic hierarchy process (AHP) to determine the weights of the attributes. Thus, the modified FMADM is a hybrid of knowledge-driven and data-driven models. The FMADM is demonstrated on a dataset taken from literature. The dataset consists of 364 web services whose quality is described by 9 attributes. Here, the attributes are treated as criteria, which are fuzzy sets and web services as alternatives. In this paper, min operator, compensatory and operator and product operator are used in aggregating the nine attributes while computing final ranks of web services and their ranks are compared. From the experiments, we conclude that min operator and compensatory and operator produced almost identical rankings and comparable to the classification done in literature.