Learning Determining Sets of Finite Partially Defined Functions
Dan A. Simovici, Dan Pletea and Rosanne Vetro
Determining sets for partially defined functions are sets of variables whose values determine uniquely the values of these functions. We propose two algorithms for the identification of determining sets. The first approach is inspired by the well-known Apriori algorithm for frequent item sets. The second algorithm makes use of information-theoretical techniques. A comparative and experimental evaluation of these approaches is also presented.
Keywords: determining set of variables, Rymon trees, entropy