Focusing on novel database application scenarios, where data sets arise more and more in uncertain and imprecise formats, in this paper we propose a novel decomposition framework for efficiently computing and querying multidimensional OLAP data cubes over probabilistic data, which well-capture previous kind of data. Several models and algorithms supported in our proposed framework are formally presented and described in details, based on well-understood theoretical statistical/probabilistic tools, which converge to the definition of the so-called probabilistic OLAP data cubes, the most prominent result of our research. Finally, we complete our analytical contribution by introducing an innovative Probability Distribution Function (PDF)-based approach, which makes use of well-known probabilistic estimators theory, for efficiently querying probabilistic OLAP data cubes, along with a comprehensive experimental assessment and analysis over synthetic probabilistic databases.
