Given an algorithm the quality of the output largely depends on a proper specification of the input parameters. A lot of work has been done to analyze tasks related to using a fixed model [25] and finding a good set of inputs. In this paper we present a different scenario, model building. In contrast to model usage the underlying algorithm, i.e. the underlying model, changes and therefore the associated parameters also change. Developing a new algorithm requires a particular set of parameters that, on the one hand, give access to an expected range of outputs and, on the other hand, are still interpretable. As the model is developed and parameters are added, deleted, or changed different features of the outputs are of interest. Therefore it is important to find objective measures that quantify these features. In a model building process these features are prone to change and need to be adaptable as the model changes. We discuss these problems in the application of cellPACK, a tool that generates virtual 3D cells. Our analysis is based on an output set generated by sampling the input parameter space. Hence we also present techniques and metrics to analyze an ensemble of probabilistic volumes.
Download