Automatic Acquisition of User Models of Interaction to Evaluate the Usability of Virtual Environments

摘要

Evaluation is an essential step in the research and development of software, particularly for new technologies such as Virtual Environments. One of the challenges to such evaluation is to collect data needed for analyzing the behavior of the users of the virtual environment. Conventional acquisition of evaluation data is time-consuming and prone to bias. In this paper, we present a taxonomy to assist identification and collection of appropriate variables for automatic data collection. We further show how these variables, such as navigation paths and characters met, can be used to capture the behavioral interaction of learners in a virtual ecosystem and to produce a user-model to evaluate the usability of the world under development.