Adaptive Generation of Workflow Models for Human-Computer Interaction
| Home | People | Publications |
|
The goal of this project is to develop a method for improving the efficiency of web-based interactive systems. Web-based systems are normally designed by hand using design guidelines, the structure of the information and intuitions about the intended users and how these will use the system. A version of its may be evaluated with a user panel. After some time the system may be revised because there is new content, new functionality or a new implementation environment. In this project we want to incorporate two additional resources: (a) similar systems and (b) usage data. The system that we are developing is able to compare a draft web-based system with a collection of other systems that is compiled by the user. The result of the comparison is a list of terms, pages, link structures, scripts, etc. in which the current draft system differs from systems in the list. This can then be used to refine the draft. The main difficulty here is how to compare websites and how to detect general characteristics of a collection of websites. The other resource is usage data. If a system has been used for some time, data can be collected on users workflow. We plan to combine the comparisons of the content and the structure with workflow data to propose improvements, building on methods for analysing and using workflow data that we developed in an earlier project (DUMPERS). The research is done by the Human-Computer Studies Laboratory in collaboration with Euro IT&C and supported by SenterNovem in the context of the IOPMMI program. |
