Model Composition Using Filter Spaces

Decision support systems (DSS) typically contain models and data to facilitate decision-making. Users of DSS in response to a decision making situation, often execute a sequence of models, where inputs to a model in the sequence are obtained from outputs of other models upstream in the sequence and from database retrievals. This problem of generating a sequence of models from the set of available models is known as the model composition problem. Most existing proposals on model composition do not provide effective support for integrating data from multiple data sources, and also lack the ability to use data stored in multiple databases. We introduce a new theoretical construct called filter spaces to facilitate model composition and execution. Filter spaces are n-dimensional spaces that facilitate the automation of model composition and execution. In this paper we demonstrate how filter spaces can address the major shortcomings of existing proposals on model composition. Decision support systems – Model management – Model composition