Control Representation in an EDA Assistant

To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be extracted and examined in detail. The search space of EDA operations is enormous, too large to be explored directly in a data-driven manner. More abstract EDA procedures can be captured, however, by representations commonly used in AI planning systems. We describe an implemented planning representation for Aide, an automated EDA assistant, with a focus on control issues.