New Algorithm for Computing the Core Based on Information Entropy

Discernibility matrix method is an important method to design algorithm for computing the core based on information entropy. In this method, the core is found by discovering all discernibility elements of discernibility matrix. So this method is very time consuming. To improve the efficient of computing the core based on information entropy, the core of the simplified decision which is the same as the core of the old decision table is found. The simplified decision table has a new property which the objects with condition attribute value of simplified decision table are order when they are looked as numbers. The core based on information entropy can be found to use the order only by searching a small quantity of discernibility elements of the simplified discernibility matrix. On this condition, an efficient algorithm for computing core based on information entropy is designed. And it´s time complexity is O(|C| |U|)+O(|C|²|U/C|). Finally, the example and experiment results show the efficiency of the new algorithm.