Knowledge-based data generation

Data analysts may not always have the data they need to effectively carry out the analysis task and, in certain circumstances, these data can only be derived or generated using existing information. In this paper, we describe a problem solving framework which involves data generation as an important component and present a complex real-world application involving mass spectral data analysis where a substantial amount of domain knowledge is required to infer an essential set of data crucial for achieving the application task. We show how the framework is applied to achieve the analysis task and how relevant domain knowledge was acquired, represented and applied to generate the data set. Finally the work is evaluated and its implications are discussed.