Characteristic matrixes-based knowledge reduction in dynamic covering decision information systems

In practical situations, dynamic covering decision information systems that change over time are of interest because databases of this kind are frequently encountered. Incremental approaches are effective in performing dynamic learning tasks because they can make the best use of previous knowledge. In this paper, motivated by the need for knowledge reduction of dynamic covering decision information systems due to variations in the object sets, we present incremental approaches for computing type-1 and type-2 characteristic matrixes of dynamic coverings. We update the characteristic matrixes with regard to two aspects: immigration and emigration of objects. Then, we provide incremental algorithms to compute the second and sixth lower and upper approximations of sets in the dynamic covering approximation spaces. The experimental results confirm that the computational complexity of constructing approximations of concepts is significantly reduced using the incremental approaches. Finally, we perform knowledge reduction of dynamic covering decision information systems by using the incremental approaches.