Visualization and user interaction techniques
The overall objective is to fully integrate the user into the process of environmental node discovery, service orchestration and decision support. In particular, wel will target the development of visual interaction techniques for: (i) query expansion and feedback during node discovery; (ii) uncertainty metrics derivation (via visual analytics methods); (iii) service orchestration, and (iv) decision support.
Visual Analytics (VA) is a relatively new scientific discipline, which aims at integrating methods from the fields of visualization, human computer interaction, and automatic analysis techniques. The scientific area developed quickly, and has been adapted to a considerable variety of problems. All of them have in common that large amounts of potentially uncertain information are involved, to which purely automatic solutions cannot be applied since they require a human user to guide the analytic sense-making process in order to generate meaningful results.
In PESCaDO, the work on several topics will profit from VA – among others, uncertainty metric derivation, as shown in the following picture:
VA allows for intervention in any of the actions to be taken during uncertainty metric derivation; for instance, for the association of confidence or weight scores to the individual parameters considered in the metric:
The next steps regarding visualization and user interaction within PESCaDO for the end user will comprise the better support for query generation and result representation by introducing ad-hoc and combinable visualization techniques, intelligently guiding the user through the available input mechanisms, as well as provide further personalization of the UI. Also, the steering of the node discovery procedures will be enhanced through interactive feedback for the employed classifiers.
Harald Bosch <firstname.lastname@example.org>