Abstract: | This work investigates the properties of the use of dimensional analysis based causal ordering and automatic text generation for the generation of informative explanations for model based reasoning. It presents a strategy of how to generate structured text and perform content selection based on device structure. It will show that existing explanation facilities for model based reasoning can be extended using the strategy coupled with dimensional analysis based causal ordering: information about value change of variables and their causal relations can be achieved and presented without relying onother model based reasoning techniques than causal ordering. The strategy for content selection and text structuring enables the creation of informative explanations from possibly very complex devices.
|