Multi-model visualization based on integration of data models in semantic network environment

The paper considers the problem of integration of data models for visualization purpose. Integration implies that the set of the data transformations from one model to another is defined and may be realized. A set of visual elements with attached attributes is offered as the final result, this set being an abstraction of the representation of vector graphic information. Conceptual structures of models to be integrated may start acting during computation process required for getting visual representation differently. It is necessary that differently acting conceptual structures (concepts and frames) be parted or separated to provide correct visualization. To have practical separation it is suggested to include in the supporting system the component called semantic separator. In general case data separation is a cognitive task. Use of methods of applicative computational systems gives some advantages. For example, it gives good means for supporting a default mode. Prototype implementations of methods of the semantic separation were tested in developing computer aided law applications.