基于粗集理论和证据理论的多源信息融合方法.pdfVIP

第33 卷第4 期 Vol. 33, No.4 2004 年8 月 Information and Control Aug. , 2004 :2004) 1 2 2 2 黄 , 陈森发, 周振国, 张文红 ( 1. , 210007; 2. , 21009 ) : , Kohonen , . , , . , , . , . : ; ; Kohonen ; ; DS ; :TP391 :A Multisource Information Fusion Method Based on Rough Sets Theory and Evidence Theory 1 2 2 2 HUANG Kun , CHEN Senfa , ZHOU Zhenguo , ZHANG Wenhong ( 1. No. 28 Research Institute of Electronics Technology Group Corp oration of China, Nanj ing 210007, China; 2. Institute of System Engineering , Southeast Uni ersity, Nanj ing 21009 , China) Abstract: Aiming at the problem of multisource information fusion, sample information is discretized and simplified with discretization algorithm based on Kohonen network and rough sets theory, and an experiential decisionmaking table for decisionmaking problem is established. On the basis of that, using the experiential decisionmaking table and the relation ship between rough sets theory and evidence theory , basic probability assignment and condition probability assignment of the related evidence about decisionmaking information are calculated. Finally, according to the combination rule, the above condition probability assignment is combined, and decisionmaking is also done according to the rules of decisionmaking. Practical application shows that the experimental result of this method is satisfactory. Keywords: information fusion; rough sets theory; Kohonen network; discretization; DS evidence theory; character recognition (Introduction)