Model extraction using context information

This work describes a new approach for behaviour model extraction which combines static and dynamic information. We exploit context information as a way of merging these types of information. Contexts are defined by evaluated control predicates and values of attributes. They create a nested structure that can facilitate the extraction of causal relations between system actions. We show how context information can guide the process of constructing LTS models that are good approximations of the actual behaviour of the systems they describe. These models can be used for automated analysis and property verification. Augmentation of the values of attributes recorded in contexts produces further refined models and leads towards correct models. Completeness of the extracted models depends on the coverage achieved by samples of executions. Our approach is partially automated by a tool called LTSE. Results of one of our case studies are presented and discussed.