Research on Attribute Discretization for Combat Simulation Data
Ji Luo · Jisuanji fangzhen · 2013
Data attribute discretization is an important part of pre-processing methods for combat simulation data.This paper discussed the necessary for data attribute discretization and proposed a discretization algorithm of combat simulation data based on improved attribute significance and information entropy( DIAFIE). As the base of data value area partition for combat simulation data,the attribute significance was defined and the values of data set was divided into many intervals based on clustering. The discretization results quality was ensured by computing information entropy of the adjacent intervals. Experiment shows that the proposed algorithm has the advantages of less points and high accuracy of classification and is efficient to deal with attribute discretization.