The Shapley Value of Inconsistency Measures for Functional Dependencies

Quantifying the inconsistency of a database is motivated by various goals including reliability estimation for new datasets and progress indication in data cleaning. Another goal to attribute individual tuples level responsibility overall inconsistency, thereby prioritize explanation or inspection dirt. Therefore, quantification attribution have been subject much research Knowledge Representation and, more recently, Databases. As many other fields, conventional sharing mechanism Shapley value from cooperative game theory. In this paper, we carry out systematic investigation complexity common measures functional-dependency (FD) violations. For several establish full classification FD sets into tractable intractable classes with respect Shapley-value computation. We also study approximation cases.