Inferring Approximate Functional Dependencies from Example Data

This paper proposes a kind of PAC (Probably Approximately Correct) learning framework for inferring a set of functional dependencies. A simple algorithm for inferring the set of approximate functional dependencies from a subset of a full tuple set (i.e. a set of all tuples in the relation) is presented. It is shown that the upper bound of the sample complexity, which is the number of example tuples required to obtain a set of functional L:"-, dependencies whose error is at most c with a probability of at least 1 -6, is O(~/~V~), where n denotes the size of the full tuple set and the uniform distribution of examples is assumed. An experimental result, which confirms the theoretical analysis, is also presented.