An Improved Genetic Algorithm for Attribute Reduction in Rough Set Theory

Since attribute reduction is NP-hard problem, in order to obtain the minimum relative reduction effectively in decision table, an attribute reduction algorithm in rough set theory based on improved genetic algorithm is proposed. At first, the dependence concept of decision attribute on condition attribute is introduced. At the same time, the relative importance of the condition attribute with respect to decision attribute is defined. On the basis of these conditions , a relative core is computed based on attribute significance as heuristic information, which adds to the initial population so that the evolution of chromosomes is guided. What's more, the control factor is introduced in an appropriate function to ensure the stronger distinction ability. Therefore, the proposed appropriate function can reduce the computing time to obtain the best search results. To evaluate our proposed algorithm, a numerical experiment is carried out on the Zoo dataset at last. The results demonstrate that the proposed algorithm can obtain the attribute reduction more effectively than the existing algorithms.