Judgement and Analysis Rules for Ontology-driven Comparative Data Analysis in Data Warehouses

Online analytical processing tools facilitate interactive inspection of aggregated measures of groups of data in data warehouses. In comparative data analysis, business analysts assess measures of a group of interest against measures of a group of comparison. Judgement rules annotate comparisons with background knowledge otherwise tacit to the business analyst. Analysis rules specify comparisons at different granularities and result in recommendations for further analyses. Judgement and analysis rules build on multidimensional ontologies, which simplify definition, reuse, and sharing of comparisons, and on comparative scores, which make explicit the results of comparisons. In this paper, we introduce conceptual modelling of judgement and analysis rules together with their organisation and multidimensional contextualisation in rule families, explain different rule evaluation strategies, and briefly report on the implementation of the approach.