A User's Preference Based Method for Web Service Selection

摘要

Both criterion values and userȁ9;s preference of criterion should be taken into account when selecting web services with same functionalities. Usually, user describes his/her preference of criterion using qualitative concept. But the quantitative value corresponding to qualitative concept is uncertain. Thus, the subjective weight acquired based on user''s preference is uncertain. This paper provides a method to select web service using uncertain subjective weight. Firstly, we present two ways of calculating subjective weight; Normalization and Least square method. The subjective weight is described using cloud model. Secondly, the criterion values of each web service are synthesized using marginal substitution method with indifference curve of Cobb-Douglas preferences. Thirdly, an algorithm for service selection is presented. Using this scheme, an experiment is conducted on selecting weather forecast web service. The result of the experiment is validated by the sensibility analysis of weight.