Temporal Decision Trees for Diagnosis:An Extension

Model-Based Diagnosis often cannot be directly exploited for embedded software, which must run with very strict constraints on memory and time. It is however possible to compile the knowledge explicited by a model-based diagnostic engine into a decision tree that serves as the basis for on-board diagnostic software. In order to exploit temporal information possibly present in the model, temporal decision trees for diagnosis have been introduced. This paper presents an extension to the temporal decision trees framework that widens its applicability.