Runtime data-driven adaptation of composite e-Health services

In this paper, an approach to e-Health service adaptation, based on input data analysis is presented. The developed approach can be applied to various services, not only e-Health, however, the motivating scenario presents interesting aspects of the problem and a challenge for efficient service adaptation. The adaptation is presented here as an extension of service composition approach, allowing for increase of efficiency compared to the standard approach, when non-functional parameters are static and their values are not considered to change irrelevantly of the data provided on the input. The adaptation approach has been developed for the service composition platform and its execution engine, to allow for automated service adaptation, given the results of service composition. The tests have been performed using the service composition platform and classification methods from WEKA workbench software.