Attribute reduction of an information system is a key problem in
rough set theory and its applications. Using computational
intelligence (CI) tools to solve such problems has recently
fascinated many researchers. CI tools are practical and robust for
many real-world problems, and they are rapidly developed nowadays.
However, some classes of CI tools, like memory-based heuristics,
have not been involved in solving information systems and data
mining applications like other well-known CI tools of evolutionary
computing and neural networks. In this paper, we consider a
memory-based heuristic of tabu search to solve the attribute
reduction problem in rough set theory. The proposed method, called
tabu search attribute reduction (TSAR), shows promising and
competitive performance compared with some other CI tools in terms
of solution qualities. Moreover, TSAR shows a superior performance
in saving the computational costs.