Causal Explanation from Mathematical Models
نویسنده
چکیده
Decision support systems may require the use of existing complex mathematical models. It is desirable to reduce the equations of such a model to an explanatory causal form to support decision making, for example in making decisions about the greenhouse effect. This paper presents an approach to constructing causal explanation from existing complex mathematical models. The method combines information from the user and the use of heuristics to reduce the equations to a suitable form. It is then able to provide a reasonable causal explanation about the behaviour of the system modelled. The method is generic and has been tested on a wide range of domains. This method has been implemented in Hypercard on a Macintosh to provide an accessible graphically oriented decision support tool.
منابع مشابه
Modelling with Context-Dependent Causality
Decision support systems may require the use of existing complex mathematical models. It is desirable to reduce the equations of such a model to an explanatory causal form to support decision making. It is argued here that a combination of extracting information from the user and heuristics to reduce the equations to a suitable form are sufficient to produce a reasonable causal explanation from...
متن کاملMachine-generated Explanations of Engineering Models: A Compositional Modeling Approach
We describe a method for generating causal explanations, in natural language, of the simulated behavior of physical devices. The method is implemented in DME, a system that helps formulate mathematical simulation models from a library of model fragments using a Compositional Modeling approach. Because explanations are generated from models that are dynamically constructed from modular pieces, s...
متن کاملContext~DependentCausal Explanations
We outline a way of generating causal explanations from mathematical models. This approach is derived from the causal ordering theory of Iwasaki and Simon (Iwasaki and Simon, 1986a, 1986b; Iwasaki, 1988). Rather than produce a single causality by propagating causality from variables whose values are determined from outside the model, we allow causality to be propagated from variables within the...
متن کاملInductive Learning, Uncertainty and the Acquisition of Causal Models
Causal models can be regarded as fundamental knowledge bases consisting of rules for generating explanations and predictions. As we all know, such inferences are not free from uncertainty. In an experiment about acquiring causal models by induction, we investigate the impact of the validity of such models on the certainty of inferences. The results indicate that preliminary models are revised i...
متن کاملOn the Limits and Possibilities of Causal Explanation with Game Theoretical Models: The Case of Two Party Competition
Regression-based pathand structural equation-models have two major drawbacks, if they are used for the causal explanation of social phenomena: they are too deterministic and neglect the intentions of the concerned actors as a central driving force of the analysed processes. In order to explain the distribution-effects of two party competition, this article proposes an alternative modelling appr...
متن کامل