An Incremental Feature Subset Selection Algorithm Based on Boolean Matrix in Decision System

Abstract

At present, some static feature subset selection algorithms are computationally time-consuming when decision table often change, facing such situations, in fact, the feature subset selection can be computed by utilizing the original decision table. Thus an incremental computation feature subset selection algorithm is provided in this paper, which originated from Boolean matrix technique, to select useful features for the given data objective efficiently. At first, the simplified decision table is introduced. Then using the simplified decision table to design Boolean matrix, the simplified Boolean matrix is constructed. Meanwhile, it is proved that the feature subset selection of the simplified Boolean matrix as the same to that of Boolean matrix. At last, the results on an example show that the efficiency and feasibility of the algorithm.