Discretization Algorithm Based on Cut Set Knowledge Granularity

NP-hard problem has been proved for searching the minimum cut set in discretization. Adopt the concept of cut knowledge grannulation, give the definition of cut´s important degree, using cut´s important degree as heuristic information, propose the algorithm of discretization based on cut knowledge granularity (DACKG) .First, adding cut core in empty cut set, then individually selecting the important cut to add, until the knowledge granularity of cut set is equal to the knowledge granularity of the candidate cut set, then the cut set is the system of minimum cut set. Theoretical analysis shows that the algorithm´s time complexity is polynomial. Four number of data set is applied to test the performance of the algorithm and the experimental result is compared with that of other discretization algorithm. The example proved this algorithm is effective.