A Parallel Discrete Algorithm Based on Breakpoints Information Entropy Clustering
Equivalence relation in rough set-based model extraction rules, the continuous condition attribute were discretized, but so far, no algorithm for an efficient discrete enough to handle large date sets. Firstly this paper presents a two-step discrtization algorithm based on breakpoint information entropy clustering, clustering algorithm based on information entropy was used to reduce the size of candidate breakpoints, and then entropy-based discretization algorithm for final discretization to obtain die results break point set; Secondly, parallel calculated based on Hie two-step breakpoint cluster entropy discretization algorithm for further improvement, the breakpoint information entropy clustering discrete algorithm is given based on parallel discrete; Finally simulation experiments demonstrate the efficiency of the algorithm.
rough set discretization information entropy clustering parallel computing
LIU Jing LIU Jing-bo LU0 Wei-ming
Experimental Center, Chongqing Three Gorges University WanZhou Chongaing, China College of Mathematics&Computer Science Chongqing Three Gorges University WanZhou Chongqing, China
国际会议
重庆
英文
477-481
2011-08-20（万方平台首次上网日期，不代表论文的发表时间）